adviser_search
ChatGPTSearch SEC-registered investment advisers (RIAs + Exempt Reporting Advisers) by firm name. Returns CRD number, type, regulatory assets under management, state, and whether the firm has Form ADV Item 11 disciplinary disclosures. ~23.4k advisers covered.
advisers_flagged
ChatGPTSEC-registered investment advisers reporting Form ADV Item 11 disciplinary disclosures (criminal / regulatory / civil), ranked by disclosure count. ~2,291 advisers flagged. Filter by state, type (RIA/ERA), or kind (criminal/regulatory/civil).
ag_census_county_search
ChatGPTSearch USDA Census of Agriculture by county/state name. Returns the place_id ({state_fips}-{county_code}) to pull a county's farm economy — number of farms, land in farms, market value of products sold, net farm income, and top commodities by sales (2022). Aggregate federal statistic — zero PII.
agent_access
ChatGPTRows and headlines of Closed to Machines, a MORVS register of which public institutions serve an automated client that names itself and gives a contact address, what their robots rules say for each path checked, and where the official route to their data is (US federal agency hosts from the Federal Register's agency list, the 50 state legislatures, the G20 members' national government portals and national law sites, and MORVS's own two hosts; checked weekly by two clients, curl and Python urllib, from one residential US vantage, both sending the same identified user agent). Optional filters: jurisdiction (a jurisdiction or ISO 3166 code the register carries, such as Illinois, US-IL or EU), slice (a slice ID the register carries, such as federal, legislature, g20_portal, g20_law or self), status (a result class ID the register carries, such as refused_held, mixed, refused_unconfirmed, robots_rule_stop, no_answer, served, directory_link_dead or not_checked_by_policy), vendor (an edge responder the register names from response headers, such as Akamai or Cloudflare), token (a product token the register evaluates; each row then shows its home-path verdict for that token) and has_route (true or false). With filters: counts by class (and by home-path verdict for the token) over all matching rows, and up to 25 rows (institution, host, slice, jurisdiction, class, status label and streak, robots state, the home-path verdict, responders, route review state, and the arguments that return the full row). With no inputs: the register's headlines with their basis, the counts by class for each slice, the result-class definitions, and the coverage gaps (rows not checked by policy, dead directory links, rows without a robots reading, and rows with no route review). Nothing here grants permission: a robots rule is a request a site publishes, not a licence, and a page that answered certifies no deeper path. A refusal is the edge's response to this client at a stated time; the vendor named is the responder seen in headers, not the institution's policy, and no record shows a reason. "No answer" (DNS, connection, TLS or timeout) is never a refusal; none_found (reviewed, nothing found where listed) and not_reviewed are kept apart, and neither shows that no route exists. Each answer leads with uncertainty[], then as_of (the cycle date), vantage, the source page on morvs.ai and the licence, and carries the question, vantage, label, robots-not-licence and absence notes (and the self-row note for MORVS's own hosts) and the claim rules. Data more than 9 days old makes each robots verdict and next step unknown, with a note. A failed or malformed read of the register is reported as temporarily unavailable, never as an empty answer. Answers are kept under about 25 KB. Organisations only; officials appear only by office and no person is named.
agent_access_changes
ChatGPTThe change log of Closed to Machines, a MORVS register of which public institutions serve an automated client that names itself and gives a contact address, what their robots rules say for each path checked, and where the official route to their data is (US federal agency hosts from the Federal Register's agency list, the 50 state legislatures, the G20 members' national government portals and national law sites, and MORVS's own two hosts; checked weekly by two clients, curl and Python urllib, from one residential US vantage, both sending the same identified user agent), from since (a date, YYYY-MM-DD, on or after which entries are returned), with optional host (a host the register covers) and kind (a change kind ID, such as register_created). The log is append-only from the build-day baseline (register_created) and never backfilled: robots-file changes with the home-path verdicts that moved, crawl-delay and Content-Signal changes, home redirect changes, route answer changes, rows added, and a class change once the new class has held three weekly cycles. Returns up to 25 entries, newest leading, or "no change recorded since <date>" with the kinds recorded so far. Nothing here grants permission: a robots rule is a request a site publishes, not a licence, and a page that answered certifies no deeper path. A refusal is the edge's response to this client at a stated time; the vendor named is the responder seen in headers, not the institution's policy, and no record shows a reason. "No answer" (DNS, connection, TLS or timeout) is never a refusal; none_found (reviewed, nothing found where listed) and not_reviewed are kept apart, and neither shows that no route exists. Each answer leads with uncertainty[], then as_of (the cycle date), vantage, the source page on morvs.ai and the licence, and carries the question, vantage, label, robots-not-licence and absence notes (and the self-row note for MORVS's own hosts) and the claim rules. Data more than 9 days old makes each robots verdict and next step unknown, with a note. A failed or malformed read of the register is reported as temporarily unavailable, never as an empty answer. Answers are kept under about 25 KB. Organisations only; officials appear only by office and no person is named.
agent_access_check_url
ChatGPTWhat a public institution's published robots rules say for one URL and one product token, from Closed to Machines, a MORVS register of which public institutions serve an automated client that names itself and gives a contact address, what their robots rules say for each path checked, and where the official route to their data is (US federal agency hosts from the Federal Register's agency list, the 50 state legislatures, the G20 members' national government portals and national law sites, and MORVS's own two hosts; checked weekly by two clients, curl and Python urllib, from one residential US vantage, both sending the same identified user agent). Takes url (http or https) and optional product_token (a robots.txt product token such as GPTBot, ClaudeBot or Googlebot; MORVS when omitted). The rules are evaluated under RFC 9309 against the robots.txt text the register archived in its latest cycle, at the exact path and query of the URL; this server never requests the site. Returns robots.result (disallow_rule_matches, allow_rule_matches, no_rule_matches, no_rules_html_served, no_rules_rfc_unavailable, complete_disallow_rfc_unreachable, undetermined_edge_refused, or unknown), the decisive rule with its line, crawl_delay_s and max_requests_per_day where the file sets a delay for the token, the edge observation (the row's result class, status label, streak_weeks and each dated check with its responder), the official machine route where one was reviewed (URL, type, when it answered without a key, the route path's own robots result for the token, terms URL and status) or none_found or not_reviewed, and next_step (use_official_route, stop, no_rule_against_path_check_terms or unknown). The same host gives different verdicts for different paths. A host the register does not cover returns unknown with the absence note. Nothing here grants permission: a robots rule is a request a site publishes, not a licence, and a page that answered certifies no deeper path. A refusal is the edge's response to this client at a stated time; the vendor named is the responder seen in headers, not the institution's policy, and no record shows a reason. "No answer" (DNS, connection, TLS or timeout) is never a refusal; none_found (reviewed, nothing found where listed) and not_reviewed are kept apart, and neither shows that no route exists. Each answer leads with uncertainty[], then as_of (the cycle date), vantage, the source page on morvs.ai and the licence, and carries the question, vantage, label, robots-not-licence and absence notes (and the self-row note for MORVS's own hosts) and the claim rules. Data more than 9 days old makes each robots verdict and next step unknown, with a note. A failed or malformed read of the register is reported as temporarily unavailable, never as an empty answer. Answers are kept under about 25 KB. Organisations only; officials appear only by office and no person is named.
agent_access_institution
ChatGPTOne row of Closed to Machines, a MORVS register of which public institutions serve an automated client that names itself and gives a contact address, what their robots rules say for each path checked, and where the official route to their data is (US federal agency hosts from the Federal Register's agency list, the 50 state legislatures, the G20 members' national government portals and national law sites, and MORVS's own two hosts; checked weekly by two clients, curl and Python urllib, from one residential US vantage, both sending the same identified user agent), in full, found by query (a host, an institution's name as listed, a jurisdiction, an ISO 3166 code, or a row ID such as federal:www.usda.gov; when several match, each in short form with the arguments that return it). Returns the institution, its slice and authority (its Federal Register agency page or MORVS's curated list), the CISA .gov registry match for its domain (domain type, organisation and suborganisation; never a contact), its result class with its definition, status label, streak and class history, the robots.txt facts (status, archive hash, crawl delays, sitemaps, Content-Signal, intent-ambiguity lines and route pointers), the home-path verdict for each product token the register evaluates, grouped as training, search or user-acting (with each token's operator and whether its group follows the operator's own description), the dated checks with their responders, the route review and routes with their terms, the change-log entries for the host, and GSA Site Scanning's record for a federal host; the redacted robots excerpt when it fits. Nothing here grants permission: a robots rule is a request a site publishes, not a licence, and a page that answered certifies no deeper path. A refusal is the edge's response to this client at a stated time; the vendor named is the responder seen in headers, not the institution's policy, and no record shows a reason. "No answer" (DNS, connection, TLS or timeout) is never a refusal; none_found (reviewed, nothing found where listed) and not_reviewed are kept apart, and neither shows that no route exists. Each answer leads with uncertainty[], then as_of (the cycle date), vantage, the source page on morvs.ai and the licence, and carries the question, vantage, label, robots-not-licence and absence notes (and the self-row note for MORVS's own hosts) and the claim rules. Data more than 9 days old makes each robots verdict and next step unknown, with a note. A failed or malformed read of the register is reported as temporarily unavailable, never as an empty answer. Answers are kept under about 25 KB. Organisations only; officials appear only by office and no person is named.
agent_access_route
ChatGPTOfficial machine routes to a public institution's data (an API, bulk files, a DCAT data.json catalog, a sitemap, a feed or a download page) as reviewed in Closed to Machines, a MORVS register of which public institutions serve an automated client that names itself and gives a contact address, what their robots rules say for each path checked, and where the official route to their data is (US federal agency hosts from the Federal Register's agency list, the 50 state legislatures, the G20 members' national government portals and national law sites, and MORVS's own two hosts; checked weekly by two clients, curl and Python urllib, from one residential US vantage, both sending the same identified user agent). Found by institution_or_host (a host, an institution's name as listed, a jurisdiction, an ISO 3166 code such as US-IL or EU, or a row ID such as federal:www.usda.gov), with optional record_type (a route type the register carries). For each matching row (up to 25): its route review state, kept apart (route: a reviewed route is listed; none_found: reviewed, with the places searched; not_reviewed: no review yet), and for each route its URL and type, the evidence that ties it to the institution (kind, URL, and whether it was re-checked this cycle), the one request's record, when it answered without a key and what that answer confirms, its terms URL (or unknown), its official domain, and the route path's own robots result for MORVS and by token group, evaluated under RFC 9309 from the route host's archived robots.txt. A route that answered once without a key does not show that the data works or may be reused. Nothing here grants permission: a robots rule is a request a site publishes, not a licence, and a page that answered certifies no deeper path. A refusal is the edge's response to this client at a stated time; the vendor named is the responder seen in headers, not the institution's policy, and no record shows a reason. "No answer" (DNS, connection, TLS or timeout) is never a refusal; none_found (reviewed, nothing found where listed) and not_reviewed are kept apart, and neither shows that no route exists. Each answer leads with uncertainty[], then as_of (the cycle date), vantage, the source page on morvs.ai and the licence, and carries the question, vantage, label, robots-not-licence and absence notes (and the self-row note for MORVS's own hosts) and the claim rules. Data more than 9 days old makes each robots verdict and next step unknown, with a note. A failed or malformed read of the register is reported as temporarily unavailable, never as an empty answer. Answers are kept under about 25 KB. Organisations only; officials appear only by office and no person is named.
arms_sales_buyer
ChatGPTLook up a buyer government in the Arms Sales Register by a substring of its name, of the purchaser as the notices print it, or of a country string as DOJ prints it in its FARA reports (for example "KOREA, SOUTH"). Returns up to 5 buyers, each with its notices and estimated value notified by year delivered, its rank by estimated value, the purchaser strings as printed, its 5 largest notices by estimated value with their items as printed, and any notice of its that is outside the totals; and, side by side and never added, the FARA money for each DOJ country string mapped to it with that string's own DOJ rank (a buyer mapped to several strings, such as the United Kingdom, shows each string's rank and no single rank of its own), the government and agency principal lines, and the firms registered under FARA for a principal there that also filed 2025 lobbying reports for a named contractor family, the two registrations shown side by side and never sequenced with any notice. FARA money is DOJ's figure for every principal under the country string DOJ prints (companies, parties and media as well as governments), for six-month periods ending in 2025, given with DOJ's footnote that a country or location is not reflective of U.S. foreign policy and does not imply a recognition of sovereignty; registration is a legal disclosure. Each row is a proposed sale notified to Congress, and its value is the estimated value the notice prints: never a price, a signed case, a contract or a payment. The estimated value notified, FARA money, connected-PAC money, lobbying amounts and the inaugural and super PAC records are separate records, shown side by side: never added together and never put in date sequence with a notice. Every answer carries the register's question note: these are proposed sales the Defense Department notified to Congress, as printed in the Federal Register; a notification is not a completed sale, and the records do not show why any sale was proposed or whether it went ahead. Every answer also carries the side-by-side caption, the as-of date, the date the Federal Register copies are printed through, each file's date, the source page on morvs.ai and the licence. Institutions only: no official, signer, lobbyist, treasurer, candidate or member of Congress is named, and single-person firms are withheld and appear only as counts. A name that matches only a withheld string answers exactly like a name that matches nothing. No party label is given and no buyer government is characterised.
arms_sales_contractor
ChatGPTLook up a contractor family in the Arms Sales Register by a substring of its reviewed family name or of a company string as the notices print it. Returns up to 5 families, each with the companies the family includes and the strings as printed; the notices naming it by year delivered and by its leading buyers; the estimated value of the notices naming it (shared, never split), with the notices naming it alone, among the companies a competition will choose from, and since January 20, 2025; the joint ventures it is a parent of; and the side columns, each from its own record: its connected PAC's contributions to federal candidates' committees and its contributions to all committees, by FEC cycle (FEC bulk files; no committee it gave to, candidate or party is named), its 2025 lobbying reports (lda.gov; in-house expenses and outside firms' income are each the report's total for all issues and are never added together), the 2025 inaugural committee's report and the reviewed super PAC donors, with the caption. An empty column is a search result, not a fact about the company. Each row is a proposed sale notified to Congress, and its value is the estimated value the notice prints: never a price, a signed case, a contract or a payment. A company appears only as a notice names it, resolved to the register's reviewed family table: the value of the notices naming a family is shared, never split between the companies a notice names, never added across families and never described as company revenue. The estimated value notified, FARA money, connected-PAC money, lobbying amounts and the inaugural and super PAC records are separate records, shown side by side: never added together and never put in date sequence with a notice. Every answer carries the register's question note: these are proposed sales the Defense Department notified to Congress, as printed in the Federal Register; a notification is not a completed sale, and the records do not show why any sale was proposed or whether it went ahead. Every answer also carries the side-by-side caption, the as-of date, the date the Federal Register copies are printed through, each file's date, the source page on morvs.ai and the licence. Where FEC figures appear, the answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)). Institutions only: no official, signer, lobbyist, treasurer, candidate or member of Congress is named, and single-person firms are withheld and appear only as counts. A name that matches only a withheld string answers exactly like a name that matches nothing. No party label is given and no buyer government is characterised.
arms_sales_notices
ChatGPTList proposed sales notified to Congress in the Arms Sales Register, filtered by buyer (a substring of the buyer's name or of the purchaser as printed), year delivered to Congress (from 2017, a year the register carries), contractor family (a substring of a reviewed family name or of a company string as the notices print it) and type (notice, the default, for 36(b)(1) notices, or enhancement for 36(b)(5)(C) reports; the contractor filter applies to notices only); at most 25 rows, the most recently delivered at the top, with the full count. Each notice gives its transmittal number, the purchaser as printed, the dates delivered to Congress and published, the estimated value as printed (the major defense equipment, other and total lines), the items as printed, the principal-contractor sentence and the reviewed families it names, the sales-commission and offsets lines as printed, any emergency sentence quoted where the notice prints one, the corrections applied, and the Federal Register citation and URL; each report gives its original transmittal, its description as printed and any revised or new-items value it prints. A sentence that names a company string still under review is held and not shown. Without a contractor filter the answer gives the count and estimated value of the matching notices; with one, each matched family's count and the value of the matching notices naming it, never added across families. The notices printed with no delivery date and the notice DoD withdrew are listed, each labelled as outside every total. Each row is a proposed sale notified to Congress, and its value is the estimated value the notice prints: never a price, a signed case, a contract or a payment. A company appears only as a notice names it, resolved to the register's reviewed family table: the value of the notices naming a family is shared, never split between the companies a notice names, never added across families and never described as company revenue. 36(b)(5)(C) enhancement or upgrade reports are a separate record type, counted apart; their values are never added to 36(b)(1) values. Every answer carries the register's question note: these are proposed sales the Defense Department notified to Congress, as printed in the Federal Register; a notification is not a completed sale, and the records do not show why any sale was proposed or whether it went ahead. Every answer also carries the side-by-side caption, the as-of date, the date the Federal Register copies are printed through, each file's date, the source page on morvs.ai and the licence. Institutions only: no official, signer, lobbyist, treasurer, candidate or member of Congress is named, and single-person firms are withheld and appear only as counts. No party label is given and no buyer government is characterised.
arms_sales_summary
ChatGPTTotals from the Arms Sales Register, built from the Federal Register copies of the Defense Department's notifications to Congress under section 36(b)(1) of the Arms Export Control Act, delivered from January 1, 2017. Returns the total estimated value of the proposed sales notified to Congress with its basis and the date the copies are printed through; notices and estimated value by year delivered, with the 36(b)(5)(C) enhancement or upgrade reports counted apart; the 10 largest buyers by estimated value notified; the equal one-year windows around January 20, 2025 with the Gulf governments' share, the last window labelled open, as the Federal Register prints notices months after delivery; the 10 contractor families named in the most notices, with the value of the notices naming each; the counter-findings (the largest buyer's FARA rank, no named family among the reviewed super PAC donors, and every printed sales-commission line saying none); the transmittal-number gaps and the publication lag by year, with the Taiwan line; the FMS trust fund account as context only; and the register's headlines with their basis. The notices printed with no delivery date and the notice DoD withdrew are listed, each labelled as outside every total. Each row is a proposed sale notified to Congress, and its value is the estimated value the notice prints: never a price, a signed case, a contract or a payment. A company appears only as a notice names it, resolved to the register's reviewed family table: the value of the notices naming a family is shared, never split between the companies a notice names, never added across families and never described as company revenue. 36(b)(5)(C) enhancement or upgrade reports are a separate record type, counted apart; their values are never added to 36(b)(1) values. The estimated value notified, FARA money, connected-PAC money, lobbying amounts and the inaugural and super PAC records are separate records, shown side by side: never added together and never put in date sequence with a notice. Every answer carries the register's question note: these are proposed sales the Defense Department notified to Congress, as printed in the Federal Register; a notification is not a completed sale, and the records do not show why any sale was proposed or whether it went ahead. Every answer also carries the side-by-side caption, the as-of date, the date the Federal Register copies are printed through, each file's date, the source page on morvs.ai and the licence. Where FEC figures appear, the answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)). Institutions only: no official, signer, lobbyist, treasurer, candidate or member of Congress is named, and single-person firms are withheld and appear only as counts. No party label is given and no buyer government is characterised.
border_force_area
ChatGPTLook up a CBP area of responsibility in the Border Force Register, built from CBP's own monthly use-of-force and assault files (FY2021 on), by sector or field-office name substring. An area is a CBP office's area or unit (a Border Patrol sector, an OFO field office or a national unit), never a city or an operation; CBP's own service-area statement is quoted where CBP gives one. Returns yearly figures, or monthly figures for one fiscal_year: use-of-force incidents with their force types (force totals include vehicle and vessel force, such as tire-deflation devices), officers/agents using force, assault incidents, and officers/agents assaulted. Officers assaulted and officers using force are counted per incident: one officer in two incidents counts twice, and neither is a count of people injured. The current fiscal year is labelled partial. No rates (nothing is divided by encounters or staffing) and no causation: CBP's files do not say why counts change. A non-match says only that the name is not in CBP's files.
border_force_national
ChatGPTNational yearly totals from CBP's own use-of-force and assault files in the Border Force Register (FY2021 on; the current fiscal year is labelled partial): use-of-force incidents by force type (vehicle and vessel force included), officers/agents using force and officers/agents assaulted (both counted per incident), and assault incidents, each year with the CBP file and date it comes from. With fiscal_year, adds that year's revision log entries, each labelled as published on its date: revisions are as-published versions of CBP's files, not corrections and not evidence of misconduct. Also returns the status of DHS's department-wide use-of-force tables, the series with no newer edition (including ICE's count in DHS's table), and DHS's tables beside CBP's files with the definitions note; the two use different definitions and are never added or merged. No rates and no causation.
bought_by_your_police
ChatGPTWhat city, county and state governments' own checkbooks report paid to reviewed vendors, from Bought by Your Police, a MORVS register built from 33 city, county and state governments' own open checkbooks on Socrata portals at launch, each read for its own period, with each vendor's family taken from its own federal award words or its own product page. Optional filters: government (a name, a name with its state, a state or a 2-, 5- or 7-digit FIPS code, as the register's lookup files them), vendor (a reviewed vendor's name or a payee alias as printed), family (a family ID the register carries, or its label; with lane panel, a panel family) and lane (police, panel or broad). For one government: each of its checkbooks with its period, row unit and licence, its police-lane totals by family for that period with each vendor beside, the other-department, department-not-stated and unclassified amounts apart, the panel and the broad platforms each apart, the families with no payments found in its period, and up to 25 payment rows (payee and line label as printed, department, amount, date or year or fiscal year, row unit, the row's own words, the dataset ID, the family, the lane, and a find-in-checkbook link; broad platforms give the register's aggregates instead, as it publishes no row for them). When more than one government matches (a state, or a vendor, family or lane on its own): each government's totals by family with its own period and unit, never summed, no rows, and the arguments that return its rows (for_rows); with a vendor or family, the governments with no payments found. With no filters: the headlines with their basis, the family IDs and labels with their counts of governments, and the governments covered with their type and checkbook period. Three columns are kept apart and never added together: the police lane (payments by law-enforcement departments to vendors of the register's families, such as phone and device extraction, plate readers, gunshot detection, face recognition and drones), the records-about-people panel (what any department pays data companies such as the Equifax unit that runs The Work Number), and the broad platforms (Axon and Motorola Solutions, never in a family or any total). Payments by other departments, and rows whose checkbook names no department, are shown apart from the police lane. Figures are what the checkbook reports paid, each with its dataset, period and row unit; a payment is not a contract value, a deployment, a use or a measure of effectiveness, and what a row bought is named only where its own words say so. A vendor is described only as a vendor of its family, per its cited basis. Figures are never added across governments, whose periods and units differ; counts of governments come with each one's period. An empty result returns the register's absence note: no payments found in this checkbook's period under the reviewed names, which does not show that a government never paid that vendor. A reviewed universe, not a national census. Answers are kept under about 25 KB. The answer carries as_of, retrieved, the file dates, the question, universe, period, unit, lane, panel, broad and absence notes, the claim rules, the licence (each checkbook keeps its own) and the source page on morvs.ai. Payees come only from the reviewed allow-list of organisations; rows whose words could name a person are dropped by the register, and structured fields it withholds are marked withheld.
bought_by_your_police_vendor
ChatGPTOne reviewed vendor in Bought by Your Police, a MORVS register built from 33 city, county and state governments' own open checkbooks on Socrata portals at launch, each read for its own period, with each vendor's family taken from its own federal award words or its own product page, found by vendor (its name, or a payee alias as a checkbook prints it). Returns the reviewed entry (its column: a family, the panel or a broad platform; the family and its label, whether the family is established, the register's wording for it, each basis with its kind, URL, the award words or quoted page text, and the register's check of it with its status and date where it records one, the row rule that assigns rows by their own words, the aliases reviewed and the rows published) and the governments whose checkbooks report payments to it, each with its figures by lane and family (police lane, other departments, department not stated, unclassified, the panel, or the broad platforms) and each with its own period and row units, never summed across governments, with the arguments that return its rows; then the governments read with no payments found to it, each with its checkbook period. Three columns are kept apart and never added together: the police lane (payments by law-enforcement departments to vendors of the register's families, such as phone and device extraction, plate readers, gunshot detection, face recognition and drones), the records-about-people panel (what any department pays data companies such as the Equifax unit that runs The Work Number), and the broad platforms (Axon and Motorola Solutions, never in a family or any total). Payments by other departments, and rows whose checkbook names no department, are shown apart from the police lane. Figures are what the checkbook reports paid, each with its dataset, period and row unit; a payment is not a contract value, a deployment, a use or a measure of effectiveness, and what a row bought is named only where its own words say so. A vendor is described only as a vendor of its family, per its cited basis. Figures are never added across governments, whose periods and units differ; counts of governments come with each one's period. An empty result returns the register's absence note: no payments found in this checkbook's period under the reviewed names, which does not show that a government never paid that vendor. A reviewed universe, not a national census. Answers are kept under about 25 KB. The answer carries as_of, retrieved, the file dates, the question, universe, period, unit, lane, panel, broad and absence notes, the claim rules, the licence (each checkbook keeps its own) and the source page on morvs.ai. Payees come only from the reviewed allow-list of organisations; rows whose words could name a person are dropped by the register, and structured fields it withholds are marked withheld.
breach_notice_organization
ChatGPTFind public notifying organisations by name substring in the Washington, Delaware and California Attorney General breach-notice lists. Returns notice IDs and counts separately for each state, plus any paired Washington and Delaware notices whose notice dates are within three days. When paired discovery dates differ, the two states' records carry different dates; neither date is declared correct. Withheld names are excluded from this index and never returned or searchable. A non-match in these three states' records never means no breach: list coverage, reporting thresholds and organisation names vary. The notifying organisation may differ from the breached one.
breach_notices
ChatGPTSearch the Washington, Delaware or California Attorney General breach-notice list by public notifying-organisation name and year (notice year for Washington and Delaware, reported year for California). Timing uses only the dates in each record: Washington and Delaware results label whether the interval is beyond that state's statutory window by the record's dates. The records do not show permitted delays, including law-enforcement requests, or any enforcement outcome; the timing flag is no legal finding. Never call a notice "late" or say an organisation "violated" the law. California has no discovery date, so no timing is computed. Resident counts are notices sent to residents, not unique people, and are never added across states; Delaware also reports the company's stated total affected separately. A notifying organisation may differ from the breached one. Names that may identify individuals are withheld, never returned or searchable. Absence from these three lists is no evidence that an organisation had no breach.
broadcast_station_search
ChatGPTSearch the FCC broadcast station directory by call sign or community of license. ~109,812 US over-the-air facilities (TV/FM/AM/LPTV/translators, ~32,883 active). Returns service, channel, frequency, community, status, and license expiration. Station facts only — no ownership.
bsee_search
ChatGPTSearch BSEE offshore oil & gas operators (US Outer Continental Shelf — Gulf of Mexico, Pacific, Alaska) by name. Returns each operator's incident-enforcement totals: warnings, component shut-ins, facility shut-ins. ~674 operators, 1948-2026.
bsee_worst_operators
ChatGPTRanks offshore oil and gas operators by counts of BSEE incidents of noncompliance: facility shut-ins (default), component shut-ins, total events, or warnings. Counts come from BSEE records and are not adjusted for operator size or activity.
building_permits_search
ChatGPTSearch Census Building Permits Survey by county/state name. Returns place_id ({state_fips}-{county_fips}) to pull a county's residential-construction permit history — buildings, housing units, and valuation by structure type (1-unit / 2-units / 3-4 / 5+), 2015-2025. Aggregate federal statistic — zero PII.
cfpb_by_company
ChatGPTSearch MORVS's partial local CFPB complaint index by company. Returns indexed complaint counts, products, and narratives only where the CFPB published them. Check /coverage; this is not the full CFPB database.
chemical_accident_search
ChatGPTSearch EPA RMP hazardous-chemical facilities (Clean Air Act 112(r)) by facility or parent-company name. Returns each facility's chemical-accident count + location. ~99k facilities — the chemical-plant accident ledger EPA keeps FOIA-only (releases, fires, explosions, deaths).
chemical_accident_worst
ChatGPTRanks parent companies by harm reported in EPA Risk Management Program accident histories: deaths (default), injuries, or accident count. Counts come from RMP filings and are not adjusted for company size or number of facilities.
china_ai_filings
ChatGPTSearch the CAC public generative AI, deep-synthesis and algorithm lists by a Chinese or English substring in the listed service or algorithm name, filer legal name, curated filer English name, or CAC filing number. Registry defaults to all; province applies to generative AI rows with a CAC province field. A us_list=true filter uses confirmed tier A legal-entity crosswalks only; brand uses an exact filer legal-name allowlist. Returns the CAC number, name and filer as printed, a separately labelled own English name or MORVS translation when available, CAC notice, date and its date type. The CAC says completed filing (备案) or registered (登记); neither grants approval. Filings are neither distinct models nor distinct companies. A brand in a filer legal name does not establish ownership. No match does not establish that a service or company is absent from China.
china_ai_us_lists
ChatGPTRead the curated crosswalk from CAC filer legal names to the US Entity List and Treasury NS-CMIC list. Search by Chinese or English entity-name substring; tier defaults to A. Tier A has published evidence for a legal-entity match. Tier B is a possible match, not confirmed, and is outside headline counts. Each entry gives the list name, start date, evidence type and text, and matching CAC record IDs. The Entity List imposes export-licence requirements for specified items subject to the EAR; the NS-CMIC investment ban covers specified US-person purchases or sales of publicly traded securities. These are different restrictions, not generic sanctions. The separate US actions timeline includes the DHS DeepSeek order with Federal Register links. MORVS does not determine whether any CAC DeepSeek filer is the company named by DHS. Not being on either US list never means an entity is cleared.
clinical_trials_recent
ChatGPTRecent clinical trial activity from ClinicalTrials.gov. Filter by phase, status, and sponsor. Use for drug pipeline questions.
cluster_buys
ChatGPTCompanies where multiple distinct insiders bought open-market stock in the window. Strong conviction signal.
cluster_sells
ChatGPTCompanies where multiple distinct insiders sold open-market stock in the window. Mass-exit detector.
common_owner_rollup
ChatGPTTrace one organization owner across every Medicare facility it controls (hospital/hospice/home-health/skilled-nursing): facility_count, per-sector breakdown, PE/REIT flags, state spread. Takes a CMS owner_associate_id (from /api/v1/owners/largest or /owners/private-equity).
company_insider_pulse
ChatGPTUnified Form 4 (executed) + Form 144 (proposed-sale) insider timeline for one company.
company_profile
ChatGPTComprehensive dossier for one SEC-filing company: basic info, recent filings, insider activity windows (7/30/90d), upcoming Form 144 sales, latest financials (FY+quarterly), top 13F institutional holders, top mutual-fund holders, peers.
company_risk_flags
ChatGPTMulti-source red-flag detector for one company — restatements (8-K item 4.02), auditor changes, officer departures, insider cluster sells, upcoming Form 144 pressure, Schedule 13D filings, FDA recalls.
compare_entities
ChatGPTSide-by-side compliance-risk comparison of two entities. Runs the full 40-source screen for each and returns both risk scores, an overall verdict (who is riskier), the risk gap, and the biggest per-vertical differences. Answers 'which counterparty is riskier' in one call.
courts_by_party
ChatGPTLitigation dossier for one company: docket count, top Natures of Suit, recent cases. Answers 'is this company being sued?' and 'for what?'
courts_recent
ChatGPTRecent federal court dockets from CourtListener. Filter by party, court, or Nature of Suit (850 securities, 830 patent, 410 antitrust, 365/367 product liability, 480 consumer credit).
crop_insurance_search
ChatGPTSearch USDA RMA crop-insurance Summary of Business by county/state name. Returns the place_id ({state_fips}-{county_fips}) to pull a county's underwriting history — liability, premium, subsidy, indemnity, loss ratio, top crops (1989-2026). Aggregate federal program statistic — zero PII.
cross_vertical_active
ChatGPTDiscovery feed: list of entities with activity in 2+ regulatory verticals (SEC + FDA + contracts) inside the lookback window. Use to find entities worth pulling a full cross-vertical timeline for.
csb_investigation_search
ChatGPTSearch U.S. Chemical Safety Board (CSB) investigations of major chemical incidents — explosions, toxic releases, fires — by company, location, or keyword. ~139 independent root-cause investigations (BP Texas City, West Fertilizer, DuPont, etc.); the deep-investigation layer over EPA RMP chemical accidents.
data_broker_check
ChatGPTChecks whether a business is registered as a data broker in California, using the MORVS California Data Broker Register: the yearly registries filed with the California Privacy Protection Agency (2024 onward) and the Attorney General's 2020-2023 list, keyed by website domain so a name change at the same domain reads as one history. Give a business name, a website domain or URL, or both; matching is by registrable domain or by normalised legal name, DBA or earlier filed name (case, punctuation and legal-suffix insensitive, exact after normalisation), and when several brokers match they are returned as separate candidates, never merged. Each match carries the broker's filings as filed: what it says it collects, whom it reports having sold or shared data to, the laws it says regulate it, its own consumer-request counts, its name history and its broker page. Registration is self-reported and California-only, so a non-match is not a finding that a business is not a data broker, is exempt from registration, or has never registered. Businesses only; queries shaped like a personal name are not run.
data_broker_diff
ChatGPTCompares two files of the MORVS California Data Broker Register (a yearly CPPA registry, named by its year, or the Attorney General's 2020-2023 list) under the register's published method: a registration in the earlier file whose normalised legal name has no registration in the later file is an apparent leaver; if its website domain does have a later registration it is counted as re-registered at the same domain, with the later names, otherwise as not re-registered under either name or domain; a later registration whose name and domain are both absent earlier is net new. Returns the counts and the named registrations in each group, plus the register's own published figures for consecutive files. A registration absent from a later file is not an exit, a merger or a finding: registration is self-reported and California-only, and a business may have stopped brokering, merged, changed domain or simply not filed.
data_demand_compare
ChatGPTShows the latest headline row reported by each publisher in the MORVS Government Data Demand Register, side by side with each figure's label. The figures are each publisher's own counts for its own period and scope. Companies count differently; never add or rank their figures across companies. A missing figure means not reported, not zero. National-security figures remain bands. Aggregate counts cannot show whether any person's data was requested. Optionally restrict to one sector.
data_demand_source
ChatGPTReturns one publisher's figures for one report in the MORVS Government Data Demand Register. Figures are that publisher's own counts for its own period and scope. Companies count differently. Never add or rank figures across companies. A missing figure means not reported, not zero. National-security figures remain bands. These aggregate counts cannot show whether any person's data was requested. Defaults to the latest complete report; optional filters select exact fact fields.
dataset_freshness
ChatGPTPer-dataset freshness across the federal data hub: FRESH/AGING/STALE/UNKNOWN vs each source's expected cadence, with age in days and separate last-successful-source-observation and indexed-content watermarks. Each row states which timestamp drives status. Optional status filter (FRESH/AGING/STALE/UNKNOWN).
debarments_check
ChatGPTScreen a name against US federal contractor debarments (SAM Exclusions, ~10k entries) AND HHS OIG healthcare provider exclusions (LEIE, ~83k entries). Returns hit=true if any match. Use this for KYC + counterparty due diligence.
documented_vessel_search
ChatGPTSearch USCG documented vessels by name or official number. ~394,495 US-flag vessels (recreational, fishing, freight, towing, passenger). Returns service, flag, tonnage, length, build year, hull material, hailing port, and certificate-of-documentation status. Vessel/asset facts only — no owner names or addresses (zero PII).
doj_by_company
ChatGPTDOJ press release search for one company. Surfaces FCPA, antitrust, healthcare fraud, environmental crimes, and any other DOJ enforcement actions mentioning the name. Body-text substring match across 266k+ press releases since 2009.
doj_recent
ChatGPTRecent DOJ Office of Public Affairs press releases. Filter by free-text query or DOJ component (Antitrust Division, USAO - SDNY, Criminal Division, ENRD, Civil Rights, etc.).
dossier
ChatGPTCross-topic dossier for one entity — SEC filings, 13F holdings, mutual fund holdings, FDA recalls as firm, federal register agency mentions, federal contracts as recipient.
drone_company_check
ChatGPTChecks whether a company or agency is named in the MORVS Drone Register, which compiles US government drone approval, restriction and compliance records: the DCMA Blue UAS Cleared List and Framework, FCC UAS Conditional Approvals, FAA Remote ID Declarations of Compliance, and the drone-linked entries of the Section 1260H list, the BIS Entity List and OFAC's NS-CMIC list. Matches by name or alias (case, punctuation and legal-suffix insensitive; substrings allowed) and returns each match's designations with the government record that names it, attached reference records, and federal contract totals where loaded. A non-match is reported as not found in the register; it is not clearance or a compliance determination.
drone_product_lookup
ChatGPTFinds drone platforms, components and software by model, product or category text across the DCMA Blue UAS Cleared List and Framework, FCC UAS Conditional Approvals and FAA Remote ID Declarations of Compliance in the MORVS Drone Register. Returns each matching designation with the company, list, action, date and source record, plus the current FCC Covered List exemption context taken from the register's reference records. An FAA declaration is the manufacturer's own statement, not an approval.
drug_label
ChatGPTLook up an FDA drug label by brand, generic, or substance name. Returns indications, dosage, warnings, adverse reactions, drug interactions.
election_equipment_jurisdiction
ChatGPTFind a jurisdiction by two-letter state code and literal jurisdiction name substring. Returns coded equipment answers self-reported in the EAC's 2024 Election Administration and Voting Survey (November 2024), with make/model and MORVS-curated vendor labels. Missing and coded Other answers do not mean no equipment. The survey does not establish equipment used on 2026-11-03; check Verified Voting and the state election office. Vendor counts overlap and must not be summed. Say voters participated in jurisdictions that reported a vendor's equipment; never say that vendor counted those votes.
election_voting_systems
ChatGPTFilter the EAC's certified, under-test, Withdrawn or Decertified Systems, terminated and ESTEP e-poll-book rows by list, MORVS-curated vendor label, testing standard or proposed end-of-life notice flag. Defaults to the certified voting-system list. Returns EAC names, standards, dates and source links. A system named in the EAC's April 2026 notice remains on the certified list while the notice is proposed. Federal certification is voluntary; state law decides use. Never call a system insecure or outdated; use decertified only inside the EAC's own list name. Includes the EAC's exact security statement and the November 2024 EAVS survey label.
entity_material_events
ChatGPTCross-vertical 'alert me' feed for one entity. Returns only high-signal events (importance ≥ 60): Class I FDA recalls, Schedule 13D filings, large insider sales, restatements, large federal contracts, terminated clinical trials, EPA significant-violator status.
entity_relationships
ChatGPTCompany relationship graph (shared-officer edges): other companies whose Form-4 filings name the same person(s) as this entity — board interlock / multi-company executives / activist overlap. Accepts a ticker, CIK, or name.
entity_timeline
ChatGPTCross-vertical event timeline for one real-world entity — joins SEC filings/insider/Form144/13D + FDA recalls + FDA drug labels + federal contracts into one ordered feed. Use this when an LLM needs the FULL picture of a company across regulatory verticals. Add min_importance to filter to high-signal events only.
epa_facilities
ChatGPTEPA-regulated facilities owned by a company. Returns location, NAICS, and compliance status. Set violations=1 to only see facilities with EPA violations.
estimate_filing_dates
ChatGPTEstimates each company's next 10-Q/10-K filing date from its past SEC filing cadence. Dates are estimates, not announced filing dates.
fails_to_deliver_search
ChatGPTSearch SEC fails-to-deliver by issuer name or ticker. Returns each security's total fail shares, peak single-day fail balance, and days-with-fails over the rolling 12-month window. CUSIP joins to the SEC securities/13F-holdings layer.
fails_to_deliver_top
ChatGPTSEC fails-to-deliver leaderboard — securities ranked by total fail-to-deliver shares over the rolling 12-month window. Fails-to-deliver are a settlement-stress / naked-short red flag tracked by retail investors and market-structure researchers.
fair_market_rent_search
ChatGPTSearch HUD Fair Market Rents by area/county/state name. Returns place_id (10-digit HUD fips) to pull an area's FY2026 FMRs — the 40th-percentile gross rent by bedroom size (efficiency/0BR through 4BR). Sets Section 8 voucher payment standards. Aggregate dollar amounts — zero PII.
fcc_grantee_search
ChatGPTSearch FCC equipment-authorization grantees — the manufacturer/importer behind every FCC ID — by company name. ~50k grantees across 111 countries (China 12,962, Taiwan 3,717). A supply-chain lens that joins UFLPA forced-labor + trade-screening by name. Optional country filter.
fda_firm_dossier
ChatGPTAll FDA activity for one firm — recalls + 510(k) clearances + FAERS adverse-event reports.
fed_enforcement_recent
ChatGPTFederal Reserve banking enforcement actions (consent orders, written agreements, CMPs, removals). Optional bank filter.
federal_file_search
ChatGPTSearches the MORVS Federal Files Register of federal Privacy Act systems of records by system name, number or ID, with agency, recipient, record-category, status and exemption filters. Recipient tags or groups and record flags are pattern matches on the agency text and count at least those matches; recipient matches exclude standard agency boilerplate. DHS matches do not imply ICE; ICE has its own tag. A routine use is the agency statement that it MAY disclose records for a purpose; it does not show that a disclosure happened. An exemption claim may limit access or correction but does not decide a request. Missing systems are not proof no records are kept. Results default to active systems and include published source context.
federal_file_system
ChatGPTReturns one published federal system of records by exact Federal Files Register ID, with the agency's record description, routine-use text, exemption claim, source record and linked Federal Register notices. Standard agency boilerplate routine uses are omitted unless include_standard is true. A routine use is the agency statement that it MAY disclose; it is not evidence a disclosure occurred. A later notice may change a system without reprinting its full text, so check linked notices and effective dates. Recipient tags and record flags are text matches counted as at least; DHS does not imply ICE. Exemption claims may limit access or correction but do not decide an individual request. A missing ID does not establish that no records are kept.
flagged_carriers
ChatGPTFMCSA carriers flagged in one or more SMS BASIC safety categories (over the federal intervention threshold): Unsafe Driving, Hours-of-Service Compliance, Driver Fitness, Controlled Substances/Alcohol, Vehicle Maintenance. Returns carrier name, location, fleet size, which BASICs are flagged, and out-of-service inspection counts. ~1,526 carriers currently flagged. Filter by state, a specific BASIC, or minimum alert count.
forced_labor_company_check
ChatGPTChecks a company or facility name against the MORVS U.S. Forced Labor Company Register: every organisation named on the DHS UFLPA Entity List or in a CBP withhold release order or finding, the Consolidated Screening List (Commerce, Treasury) rows that match one of those names exactly, and issuer relationships a US-listed company states in its own SEC filing. Matching is exact on the register's name key (case, accent, punctuation and legal-suffix insensitive) or on a source-stated or reviewed alias (DHS "also known as" / "formerly known as", CBP "a/k/a"); there is no fuzzy, family-name or transliteration match, and several matching entities are returned as separate candidates, never merged. Each match carries its direct records, one per source row, with the list's own action, status and date, the source URL, the row locator and the snapshot date; with include_related (default true) it also carries filing-stated issuer relationships, which never extend an action to a parent. UFLPA listing, a CBP order or finding, and an export-control or sanctions row are separate actions under separate laws, and a non-match is not a clearance finding. Organisations only; queries shaped like a personal name are not run.
forced_labor_source_compare
ChatGPTCompares two sources of the MORVS U.S. Forced Labor Company Register: the DHS UFLPA Entity List (dhs), CBP withhold release orders and findings (cbp), and the Consolidated Screening List rows that exactly match a DHS or CBP name (csl). Returns the entity ids named in both sources under the register's reviewed name and alias joins, the ids in source_a with no reviewed match in source_b and the reverse, with counts, the register's own published overlap figures, the source rows behind each listed entity (source URL, row locator, snapshot date) and each source's snapshot date, in pages of up to 50 ids per list continued with next_cursor. Issuers linked only through a filing are never counted, so a parent never enters an overlap. The sources apply different laws; an entity with no reviewed match in the other source reads "No reviewed name or alias match in this source snapshot", which is not a clearance finding.
foreign_agents_country
ChatGPTLook up a country or location in the Foreign Agents Register by a substring of the country string exactly as DOJ prints it (as entered by registrants and never re-mapped, for example "KOREA, SOUTH"), optionally for one year (2024 or 2025; both when omitted). Returns up to 3 matching countries, each with the money agents reported receiving for its principals for the registrants' six-month periods ending in that year, as DOJ's reports print it, with DOJ's posting date; entries, entries with an amount, "None Reported" entries (never zero) and entries with no Finances line; registrants with money; the withheld part of the country total, as counts and money only; active principal rows and registrants from DOJ's bulk files; the Nature of Services of the published entries; and the largest published registrants and principals with their amounts. Every answer carries DOJ's footnote that country or location is not reflective of U.S. foreign policy and does not imply a recognition of sovereignty. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individuals are withheld: counted only, never named.
foreign_agents_enforcement
ChatGPTDOJ's own FARA administration and enforcement figures from its four semi-annual reports to Congress for 2024 and 2025, by half-year, exactly as the register publishes them: books-and-records inspections, civil causes of action, the criminal count with DOJ's own unit and measure, DOJ's enforcement sentences quoted verbatim, and the Statistical Summary (registration statements, terminations, registrants and foreign principals on file, short-form statements, agreements, fees and advisory opinions), with each report's posting date and the statistics note: DOJ's counts differ from the entries its listing prints and are shown as DOJ's, never reconciled. No person is named, and no motive or policy reason is given or implied. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". These are DOJ's counts, a separate measure from FARA receipts, LDA filing counts and PAC contributions, and never added to them.
foreign_agents_principal
ChatGPTLook up a foreign principal in the Foreign Agents Register by name substring, as filed or as rewritten to the institution or office where the filed string names a person, optionally narrowed by a substring of the country string as DOJ prints it. Returns up to 5 published principals, each with its type, its country as DOJ prints it, its active rows in DOJ's bulk files, and the published registrants representing it, with the money each reported receiving for or in the interests of that principal for the registrants' six-month periods ending in 2024 and in 2025, as DOJ's reports print it, with DOJ's posting dates. "None Reported" stays its own state, never zero. Entries by withheld registrants count only inside the country totals. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". A registration names who the agent represents, not what any official did. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individual foreign principals are withheld: counted only, never named or matched.
foreign_agents_registrant
ChatGPTLook up a registered foreign agent in the Foreign Agents Register by registrant name substring, among the published registrants (reviewed institutions active in DOJ's FARA bulk files or listed in DOJ's 2024 and 2025 reports). Returns up to 5 registrants, each with its type, registration and termination dates and status, its active foreign principals from the bulk files (published ones named, withheld ones as a count), and three federal records side by side: the money it reported receiving for its own six-month periods ending in 2024 and in 2025, as DOJ's reports print it, with DOJ's posting dates, by Nature of Services and by principal (DOJ); its Lobbying Disclosure Act filing count for 2025, or "not searched" (lda.gov); and its firm PAC's status and contributions to federal candidates' committees by cycle, with the FEC definitions (FEC). FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. The columns are never read as a sequence. A firm's PAC gives for the whole firm; nothing in the FEC records ties a contribution to any foreign client. No PAC or LDA match is a search result, not a fact about the registrant. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. Individuals and single-person firms are withheld and never named or matched.
foreign_agents_summary
ChatGPTTotals from the Foreign Agents Register, built from the Justice Department's FARA records: the money registered foreign agents reported receiving for the registrants' six-month periods ending in 2024 and in 2025, as DOJ's four semi-annual reports to Congress print it, with each report's printed total, DOJ's posting dates, the basis and the change between the two years; amounts DOJ lists as received prior to registration, kept apart and never added to a period; the share for services DOJ lists as lobbying and the full Nature of Services breakdown by year; active registrants and foreign principals by DOJ's own count for each half-year and by DOJ's daily bulk files; the register's eight headlines with their basis; the side-by-side PAC and LDA coverage counts; and the withheld counts. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". Amounts are what agents reported receiving for their own six-month periods, as DOJ prints them, never a calendar year of spending; "None Reported" is its own state, never zero. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individuals (individual registrants, single-person firms, individual foreign principals and short-form registrants) are withheld: counted only, never named.
forfeiture_agency
ChatGPTFind DOJ equitable-sharing agencies in one state or territory by agency-name substring or ORI. Returns DOJ database status and sharing labels, matched DOJ payments by federal fiscal year, and the agency's ESAC certifications by its own fiscal year with DOJ and Treasury as separate money fields. DOJ payment years and agency certification years must never be added, subtracted or treated as a gap. A certified balance is not misconduct. DOJ does not publish why an agency has a Non-Compliant or Ineligible status. A payment, review or audit finding does not establish unlawful conduct. Controlled-equipment entries itemise purchases already in ESAC spending categories and must never be added again. A seizure is not a crime or conviction. The newest ESAC year is partial; older thin years are marked in the response. Agency names are DOJ's; officials and other people are never named.
forfeiture_payments
ChatGPTSearch DOJ equitable-sharing payments to state, local and tribal agencies by federal fiscal year. Defaults to the latest published DOJ year. Filters by two-letter state, DOJ-printed recipient-name substring and minimum row amount. Returns DOJ cash, sale proceeds, total, original recipient name, match method and matched ORI when one exists. A DOJ-masked recipient stays masked; an unmatched row is still a payment. With no state filter, returns all state subtotals and a separate foreign-sharing table. Foreign sharing is never added to payments to domestic agencies. Agency ESAC certifications use each agency's own fiscal year and include Treasury sharing; never add or subtract them from these DOJ payments. A seizure is not a crime or conviction; a payment follows participation in a forfeiture, not a finding of unlawful conduct.
fund_holdings
ChatGPTWhat a mutual fund holds (NPORT-P). Pass a 10-digit CIK or series_id (S000012345).
house_price_index_search
ChatGPTSearch the FHFA House Price Index by geography name (metro, state, census division). Returns the place_id to pull a full house-price appreciation history (1975-2026, purchase-only / all-transactions / expanded-data). Aggregate federal statistic — zero PII.
inaugural_donors_donor
ChatGPTLook up an organisation donor in the Inaugural Donors Register by name substring, matched against the reviewed public name and every name as filed on the committees' Form 13 reports, optionally for one inaugural year (2025 or 2021; both when omitted, never added across years). Returns up to 5 named organisation donors, each with its reviewed type and, by year, donations accepted, refunded and net (each labelled), every donation row with its date, amount, memo and filing as filed, refunds kept apart (the filing gives no reason for any refund) and the latest filing date. For 2025 donors: the organisation's own LD-203 figure where an item was found and matched, with both amounts and, where they differ, the line that neither record explains the difference; and, for the donors in the register's side-by-side table, three more federal records beside the donation: contract obligations by fiscal year under reviewed parent records (USAspending), the connected PAC's status and its contributions to federal candidates' committees by cycle with its all-committees summary line, two measures never added (FEC), and 2025 lobbying report counts and amounts (LDA). Inaugural donations, contract obligations, connected-PAC contributions and lobbying reports are separate federal records, shown side by side and never added or put in sequence. A missing PAC, contract or lobbying record is a search result, not a fact about the organisation. Committees are named exactly as the FEC lists them, with their ids. Every answer carries the register's question note: the records describe donations reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld organisations (personal trusts, holding vehicles, businesses named for a person, names resting on initials) appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)).
inaugural_donors_summary
ChatGPTTotals from the Inaugural Donors Register, built from the Form 13 reports the 2025 presidential inaugural committee (Trump Vance Inaugural Committee, Inc., FEC C00894162) and the 2021 committee (PIC 2021, Inc., C00765040) filed with the FEC: each committee's cumulative printed totals, with donations accepted, refunded and net each labelled and kept apart, the filing ids and filing dates of the latest report and supplement and the coverage dates; organisations named against individual rows, individuals counted only; organisation money by reviewed type; the withheld categories as counts and amounts; the company families that gave in both 2021 and 2025, with their 2021 donations and share, never added across years; the LD-203 cross-check totals (registrants, items, exact, within $1 and differing counts, and each difference with both amounts and the line that neither record explains it), with the in-house registrants for whom no inaugural item was found given as totals only, never by name; the side-by-side coverage counts; and the register's headlines with their basis. Inaugural donations are not campaign contributions. Every answer carries the register's question note: the records describe donations reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Inaugural donations, contract obligations, connected-PAC contributions and lobbying reports are separate federal records, shown side by side and never added or put in sequence. Individual donors are counted, never named; withheld organisations (personal trusts, holding vehicles, businesses named for a person, names resting on initials) appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)).
income_by_zip_search
ChatGPTSearch IRS SOI income statistics by ZIP code or state. Returns the 5-digit ZIP to pull its income/tax profile — returns, AGI, average AGI per return, wages, capital gains, EITC, charitable giving by 6 AGI brackets (TY2022). Aggregate federal statistic — zero PII.
influence_chain
ChatGPTFollow-the-money chain for one entity: federal contracts + prime->sub subaward flow + registered lobbying + FEC political committees + FEC enforcement (MURs), with dollar tallies. Accepts a company name or a 12-char UEI.
judgment_fund_payee
ChatGPTFinds named organisations by payee-name substring in the MORVS Judgment Fund Register. Returns up to 50 matches by published amount, with total match count, summed bi-weekly payment-line amount, line count, agency groups, Treasury citation descriptions, and first and last report periods. Only payees confirmed as organisations appear; individuals are never named. A payment is not an admission of fault, a line is not a settlement, and the separate annual and bi-weekly Treasury products must never be added.
judgment_fund_payments
ChatGPTSearches Treasury bi-weekly Judgment Fund payment lines in one fiscal year of the MORVS Judgment Fund Register. Defaults to the latest published fiscal year, which may be incomplete. Filters by named organisation payee, Treasury responsible agency or MORVS agency group, and minimum known grouped amount. Returns separate source lines grouped by named payee, agency and report period; unnamed lines remain separate. Includes payee status, Treasury report links and source row numbers. Individuals are never named. A payment is not an admission of fault; a line is not a settlement, and split or identical lines remain separate payments. The bi-weekly and annual Treasury products are separate and must never be added.
kev_by_vendor
ChatGPTCISA Known Exploited Vulnerabilities (KEV) for one tech vendor. Returns CVE list + ransomware count + top affected products. The federal government's canonical list of CVEs being actively exploited.
kev_recent
ChatGPTRecent CISA KEV entries. Filter by vendor or ransomware-known flag.
lobbying_by_client
ChatGPTAggregate Senate Lobbying Disclosure dossier for one company: total spend, firms used, top issue codes, recent filings.
lobbying_by_issue
ChatGPTWho lobbies the federal government on a given issue. Takes an LDA issue code (e.g. HCR health, DEF defense, TAX, ENG energy, TRA transport) and returns top clients by reported spend. List codes at /api/v1/lobbying/issues.
lobbying_client
ChatGPTReturns client timelines from the MORVS Influence Register using a published key or a case-insensitive name substring. The register includes only reports with an AI, surveillance or privacy match in a lobbying-activity section; it is not a client’s complete lobbying history. Amounts are whole-report figures across all issues, never spending on a topic. In-house expenses and outside-firm income are separate and must not be added; subcontract income is excluded. Null is not zero. Filings do not establish a client’s position on a bill or a policy outcome. Returns the retrieval date and the LDA source disclaimer. A missing client is not proof of no lobbying.
mdl_lookup
ChatGPTFind a federal multidistrict litigation (MDL) in the MORVS Mass Lawsuit Register by MDL number, caption word, or company word appearing in the Panel caption. Returns the Panel caption, category, district, transfer date, pending and total actions, the last 12 monthly pending counts, and terminated entries when available. Actions are lawsuits, not people. Centralisation is not a finding on the merits. A falling pending count is not a settlement or payout. A missing match only means no matching MDL was found in the published register.
media_money_advertising
ChatGPTFederal advertising (product code R701) and public-relations (R708) contract obligations in the Media Money Register, optionally filtered by agency (a substring of the agency name or a short form such as DHS or DoD), firm (a substring of a reviewed firm's name as the award records carry it) and fiscal year (2017 to 2026). Returns, for each code, the all-agency series, the largest agencies by fiscal year with every other agency shown as the labelled other-agencies remainder, and the reviewed firms' obligations (FY2017 to FY2026, FY2025 and FY2026, as the register publishes them; every other recipient counted as other); for DHS, its R701 and R708 series, R701 by sub-agency and by reviewed firm for FY2024 to FY2026, and its seven campaign task orders with their dates, obligations and outlays as recorded and their competition fields as printed, with no subaward reported on any of them, so the records do not show which outlets carried the advertising; the Defense Department's share of FY2026 R701; and the register's FEC payee check, counted only. A firm is named only when its UEI is reviewed and its name carries a corporate form. Competition fields are shown exactly as FPDS prints them: the extent competed, the reason for other than full and open competition and the number of offers, or not printed. Each figure is named in its record's own words: contract obligations (a subscription is a purchase order, delivery order or definitive contract), grant obligations by assistance listing, Treasury account obligations, termination actions with FPDS's reason code as printed, and the laws' own text; fiscal year 2026 is as reported through the as-of date, as Defense Department contract records reach USAspending about 90 days late. Every answer carries the register's question note: these records show federal award and account obligations and the agencies' own termination records; they do not show what any outlet published, why an award was made or ended, or whether any payment affected coverage. Every answer also carries the side-by-side caption, the as-of date, each file's date, the source page on morvs.ai and the licence. Where FEC figures appear, the answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)). Institutions only: no official, consultant, journalist, treasurer or candidate is named; withheld recipients and the advertising firms whose name is a person's full name appear only as counts. A name that matches only a withheld string answers exactly like a name that matches nothing. No party label is given and no outlet's content is characterised.
media_money_family
ChatGPTLook up a reviewed recipient family in the Media Money Register by a substring of its name, of a parent company the register names, or of a name its award records carry, optionally limited to one block (A1 news publishers and wires, A2 information and data companies, A3 broadcasters and studios, B public broadcasting, C the U.S. Agency for Global Media's grantees, C2 media-development grantees). Returns up to 5 families, each with its block, reviewed UEIs and DUNS numbers, the names its award records carry, contract and grant obligations by fiscal year and by agency, its largest product codes, the split by what the awards are for (news families) or what the awards are for in the register's words (information companies, never counted as news), its termination actions by window with FPDS's reason codes, and the side columns: its connected PAC's contributions to candidates' committees and to all committees by FEC cycle (no committee it gave to, candidate or party is named), its 2025 lobbying reports (in-house expenses and outside firms' income are each the report's total for all issues and are never added together), the 2025 inaugural committee's report and the reviewed super PAC donors, with the caption. Families come from a reviewed table of identifiers, never a name search; an empty column is a search result, not a fact about the company. Recipients whose award records name a family's identifier as parent but which are not reviewed members are listed with their amounts and not counted in the family. Each figure is named in its record's own words: contract obligations (a subscription is a purchase order, delivery order or definitive contract), grant obligations by assistance listing, Treasury account obligations, termination actions with FPDS's reason code as printed, and the laws' own text; fiscal year 2026 is as reported through the as-of date, as Defense Department contract records reach USAspending about 90 days late. The blocks (news publishers and wires; information and data companies, never counted as news; broadcasters and studios; public broadcasting; the U.S. Agency for Global Media's grantees, described in the law's words; media-development grantees; federal advertising and public relations) are never added together, and a contract, a grant and a Treasury account are never added to one another. The side columns (connected-PAC money from the FEC bulk files, 2025 lobbying reports from lda.gov, the 2025 inaugural committee's report and the reviewed super PAC donors) are separate records shown side by side: never added to award money or to each other, and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal award and account obligations and the agencies' own termination records; they do not show what any outlet published, why an award was made or ended, or whether any payment affected coverage. Every answer also carries the side-by-side caption, the as-of date, each file's date, the source page on morvs.ai and the licence. Where FEC figures appear, the answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)). Institutions only: no official, consultant, journalist, treasurer or candidate is named; withheld recipients and the advertising firms whose name is a person's full name appear only as counts. A name that matches only a withheld string answers exactly like a name that matches nothing. No party label is given and no outlet's content is characterised.
media_money_summary
ChatGPTTotals from the Media Money Register, built from USAspending's award and account records, the FPDS ATOM feed, the public laws on govinfo and the Federal Register: federal money, FY2017 to FY2026, to news publishers and wires, information and data companies, broadcasters and studios, public broadcasting, the U.S. Agency for Global Media's four grantee corporations, media-development grantees and federal advertising and public-relations firms. Returns each block by fiscal year, each labelled and never summed across blocks; the news families' split by what the awards are for, from each award's own product code; the termination and cancellation actions on the news and information families' contracts, counted in two windows of equal length either side of January 20, 2025; the DHS advertising figures, with its seven campaign task orders; the Public Broadcasting Fund account (020-0151), the laws' own words on the Corporation for Public Broadcasting and CPB's own grants; the agency's account (514-0206), its grantees' grants and its wire contracts; the counter-finding; the register's headlines with their basis; and what is excluded: no savings claim, no Paycheck Protection Program loan and no CPB station grant is in any figure. Each figure is named in its record's own words: contract obligations (a subscription is a purchase order, delivery order or definitive contract), grant obligations by assistance listing, Treasury account obligations, termination actions with FPDS's reason code as printed, and the laws' own text; fiscal year 2026 is as reported through the as-of date, as Defense Department contract records reach USAspending about 90 days late. The blocks (news publishers and wires; information and data companies, never counted as news; broadcasters and studios; public broadcasting; the U.S. Agency for Global Media's grantees, described in the law's words; media-development grantees; federal advertising and public relations) are never added together, and a contract, a grant and a Treasury account are never added to one another. The side columns (connected-PAC money from the FEC bulk files, 2025 lobbying reports from lda.gov, the 2025 inaugural committee's report and the reviewed super PAC donors) are separate records shown side by side: never added to award money or to each other, and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal award and account obligations and the agencies' own termination records; they do not show what any outlet published, why an award was made or ended, or whether any payment affected coverage. Every answer also carries the side-by-side caption, the as-of date, each file's date, the source page on morvs.ai and the licence. Where FEC figures appear, the answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)). Institutions only: no official, consultant, journalist, treasurer or candidate is named; withheld recipients and the advertising firms whose name is a person's full name appear only as counts. No party label is given and no outlet's content is characterised.
media_money_terminations
ChatGPTList the termination and cancellation actions FPDS records on the news publisher and wire families' and the information companies' contracts in the Media Money Register, optionally filtered by family (a substring of a family name), agency (a substring of the agency or department as FPDS prints it, or a short form such as DHS, DoD, HHS or State) and window (before: the 645 days to January 19, 2025; since: January 20, 2025 to the as-of date); at most 25 rows, the most recently signed at the top, with the full count and the counts by window. Each row gives the family, the agency and department, the award and modification numbers, the date signed, FPDS's reason code and its description as printed (F, terminate for convenience; N, legal contract cancellation; E, terminate for default; X, terminate for cause), the amount the action itself records, any initiative code as printed and the product code. Actions are unique by agency, award and modification; a termination of a prepaid subscription can record little or nothing, and the amount is never compared with any claim. Each figure is named in its record's own words: contract obligations (a subscription is a purchase order, delivery order or definitive contract), grant obligations by assistance listing, Treasury account obligations, termination actions with FPDS's reason code as printed, and the laws' own text; fiscal year 2026 is as reported through the as-of date, as Defense Department contract records reach USAspending about 90 days late. Every answer carries the register's question note: these records show federal award and account obligations and the agencies' own termination records; they do not show what any outlet published, why an award was made or ended, or whether any payment affected coverage. Every answer also carries the side-by-side caption, the as-of date, each file's date, the source page on morvs.ai and the licence. Institutions only: no official, consultant, journalist, treasurer or candidate is named; withheld recipients and the advertising firms whose name is a person's full name appear only as counts. No party label is given and no outlet's content is characterised.
migrant_programs_program
ChatGPTOne program of the Migrant Program Money Register by key: its label, agency and award-ID rule (the FAIN prefix pattern and the listings it is booked under, or the reviewed award list for contracts) with the register's evidence; the listing note where the program's money moved between assistance listings (the unaccompanied-children program booked under 93.566 from December 2024 and re-booked older obligations in September 2026, so a series by listing alone moves with that booking), and money by listing or by kind where the register gives it; by fiscal year, net obligations, positive obligations, de-obligations, transactions, recipients and new awards, each labelled with its fiscal year and the as-of date; the 10 largest recipients each fiscal year as the register ranks them, with their class; new awards in FY2025 and FY2026 by funding opportunity; the equal twelve-month windows; class shares by year; and, for State Reception and Placement, the awards placed outside the United States that are excluded. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts.
migrant_programs_recipient
ChatGPTLook up a recipient in the Migrant Program Money Register by name substring as the award records name it, by the register's family name, or by its exact 12-character UEI, optionally limited to one program. Returns up to 5 published recipients, each with its UEI, states, reviewed class and the basis for it, and the IRS Exempt Organizations Business Master File record where one matches (EIN, the name and state the IRS lists, and the IRS's own codes); money by program and fiscal year; its 10 largest awards with obligations by fiscal year and award-level outlays as recorded, each labelled; its termination records in the agency's exact words with their dates; the subaward rows its prime awards report, as counts only (the register gives no subawardee total); and the register's side-by-side columns for its family: program money, LDA lobbying reports, connected PACs, ICE and CBP contracts and super PAC gifts, with the caption. An empty column is a search result, not a fact about the organisation. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing.
migrant_programs_summary
ChatGPTTotals from the Migrant Program Money Register, built from USAspending award records, HHS's TAGGS terminated-grants list and the IRS Exempt Organizations Business Master File: federal obligations, FY2023 to FY2026, to the organisations and governments that shelter, care for, resettle, transport and give legal help to migrants and refugees, in eight programs keyed from the award ID (ORR unaccompanied-children grants and ORR contracts, ORR refugee formula, Matching Grant and discretionary grants, State Department Reception and Placement, FEMA's Shelter and Services Program and DOJ EOIR legal access). Returns the all-programs total with its basis; money by program and fiscal year, with net obligations, positive obligations and de-obligations labelled; the equal twelve-month windows either side of January 20, 2025; recipient-class shares by program and year, with state, local and tribal governments apart from organisations and organisations split by IRS match; FEMA's Shelter and Services de-obligations of February 10 to 12, 2025 (the record gives no reason), with the government share; termination records counted by agency and record type; withheld recipients and records as counts and amounts; the exclusions, the awards placed outside the United States listed with their amounts; and the register's headlines with their basis. Awards whose record places the work outside the United States are excluded from every figure, as are FEMA's Emergency Food and Shelter Program and ICE detention money. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts.
migrant_programs_terminations
ChatGPTList the termination records in the Migrant Program Money Register, optionally filtered by agency (a name substring as the record or its program names it, or a short form: FEMA, DHS, USCIS, ICE, CBP, HHS, ACF, ORR, State, PRM, DOJ or EOIR, each resolved to the agency names the data gives; a short form the data does not name returns no rows and the answer lists the agency names in the data), program key and recipient name substring; at most 25 rows, with the full count and the count by record type. Three record types, each verbatim with its date and amount as the record gives it: rows of HHS's TAGGS terminated-grants list, in HHS's own words and columns (total obligated, expended, payment and unliquidated, labelled as HHS labels them) with the list's date; termination and recipient-request texts from the USAspending award records, with the agency's exact words, the action type as filed, the action's obligation and any executive-order tag exactly as the record shows it; and FEMA's Shelter and Services de-obligations of February 10 to 12, 2025, by recipient, each labelled the record gives no reason. A $0 termination action is not the money: any de-obligation is a separate, later action. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing.
migration_flows_search
ChatGPTSearch IRS county-to-county migration flows by county/state name. Returns place_id ({state_fips}-{county_fips}) to pull a county's net migration (returns/people/AGI in vs out, FY2022-2023) and its top origin/destination counties. Aggregate IRS statistic — zero PII.
national_security_surveillance
ChatGPTReturns the published agency counts in the MORVS Surveillance Orders Register national-security source ledger. Optionally filter by exact series ID and period. Every fact retains the agency label, source, page, quote, method and any qualifier or disagreement. Section 702 targets, FISA orders, national-security letters and border device searches use different units, laws and periods; never add them or treat them as one trend. Section 702 target counts estimate non-US persons abroad; FBI query counting changed. These are aggregate published counts and cannot identify surveillance of a person.
news
ChatGPTTime-ordered cross-topic news feed: insider trades, Form 144 notices, cluster buys/sells, Schedule 13D filings, 8-Ks, FDA recalls, weather alerts, presidential documents.
nhtsa_automation_investigations
ChatGPTSearch MORVS-reviewed NHTSA automated-driving, Level 2 driver-assistance and SGO-compliance investigations. Returns NHTSA’s structured action fields, cited SGO report IDs and published lookup links; scope and reason are MORVS-reviewed classifications, not NHTSA findings. Optionally includes linked recall campaigns with potentially affected units as filed by the manufacturer. An opened investigation or a defect petition is not a finding. A recall’s potentially affected units are not a fleet census. No NHTSA summary text or personal names are returned. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
nhtsa_by_manufacturer
ChatGPTNHTSA vehicle recall dossier for one auto OEM. Returns total recalls, top affected components, park-it count, recent recalls. Public auto OEMs only.
nhtsa_recent
ChatGPTRecent NHTSA vehicle recalls. Filter by manufacturer, make, model year. park_it=1 returns only the most urgent recalls (do-not-drive).
nih_by_org
ChatGPTNIH RePORTER dossier for one organization — total federal R&D funding received, top activity codes, recent grants. Best for academic medical centers + biotechs.
nsf_by_org
ChatGPTNSF Awards dossier for one organization — total funding, top programs, recent awards. Pairs with academic medical centers + biotech/tech startups.
patents_by_assignee
ChatGPTAggregate USPTO patent dossier for one company: total grants, top CPC classes, recent grants, inventor counts.
patents_recent
ChatGPTRecent USPTO patent grants from PatentsView. Filter by assignee company or CPC class. Use for R&D intel and IP portfolio questions.
pay_to_play_awards
ChatGPTThe largest ICE and CBP awards in the Pay-to-Play Register: awards funded by the component with a transaction in FY2025 or FY2026, ranked by Award Amount. Filter by component (ICE or CBP), fiscal year (2025 or 2026), recipient name substring, or the law File C records for the award (PL119-21, PL119-98 or account_level_only); limit 1-50 (default 20). Each award gives the recipient, its parent, the Award Amount and outlays (both lifetime figures), the award description verbatim from the award record, and the File C law link with its class (labelled or probable) or "account level only"; total and truncated give the full count. Award Amount is the award's lifetime amount, not an FY2025-26 obligation, and is never added to the fiscal-year contract totals. A File C law link records the account the award was charged to, not a judgment about the award.
pay_to_play_committees
ChatGPTList the federal candidate committees that received contributions from the connected PACs of the 100 largest ICE and CBP contractors in the Pay-to-Play Register, for one FEC cycle (2024 for 2023-24, or 2026 for 2025-26, default 2026). Filter by committee name substring, state (two letters), party (REP, DEM or other, as the FEC lists it) or company name substring; limit 1-50 (default 20). Each committee name is as the FEC lists it, with party, office, state and district from the FEC, the net amount from the mapped PACs and the split by company; total and truncated give the full count. A listing records a contribution reported to the FEC; it is not a finding about the candidate, the contribution's purpose or any vote. No individual donor is named and no candidate profile is built. PAC money is company-wide, not tied to ICE or CBP work, and comes from voluntary contributions, not company treasury or contract money; the FEC's own words are quoted.
pay_to_play_contractor
ChatGPTLook up an ICE or CBP contractor in the Pay-to-Play Register by company name substring, matched against the company, the parent as USAspending lists it, the reviewed alias note and its recipient names, among the 100 largest parents by FY2025-26 contract obligations. Returns up to 5 parents, each with three federal records side by side: ICE and CBP contract obligations by component and fiscal year (funding view, with the awarding view beside it), share, File C obligations on the border wall system lines and recipients (USAspending); connected-PAC status, committees as the FEC lists them and, by cycle, the PAC's contributions to federal candidate committees and its summary line for contributions to all committees, two measures that are never added (FEC); and lobbying reports by year, with in-house expenses and outside firms' income kept apart and never added, and IMM-coded counts (LDA). The three columns are never added together and are never a sequence. The register's note: "The register's name labels a question people ask; it is not a finding. These records do not show that any contract was exchanged for money, or that any contribution or lobbying influenced any award." PAC totals are company-wide, not tied to ICE or CBP work, and come from voluntary contributions, not company treasury or contract money; the FEC's own words are quoted. The absence of a PAC or LDA filing is a search result, not a fact about the company. A company outside the 100 largest parents is a scope statement, not an absence.
pay_to_play_totals
ChatGPTICE and CBP contract obligation totals from the Pay-to-Play Register, by component (ICE, CBP or both, default both) and fiscal year (2025, 2026, or both when fy is omitted): the funding view (contracts funded from the component's own appropriations, the headline measure) with the awarding view and the bridge between the two beside it, ICE and CBP combined in the funding view, and the share held by the 5, 10 and 25 largest parents. Also the P.L. 119-21 border wall system (Sec. 90001, CBP) by stage: appropriated, obligated and outlaid, each with its own as-of date; the two laws' lines by program activity, labelled where the program activity's name cites the law and shown as probable where only the account's availability window matches, never counted as the law's money; and the register's eight headlines with their basis and as-of dates. Appropriated, obligated and outlaid are separate stages and are never added. Contract obligations are orders placed and contracts signed, not outlays (money paid out). Every answer carries the register's question note and caption: the records are shown side by side and do not show that one caused another.
pfas_rule_impact
ChatGPTCompares two rule scenarios in the MORVS PFAS Drinking-Water Register: 2024_as_published (EPA's six 2024 PFAS drinking-water limits, final and in force) and if_proposals_final (the PFOA and PFOS limits only, as they would stand if EPA's May 2026 proposals to rescind the PFHxS, PFNA, HFPO-DA and Hazard Index limits became final; they are proposals, not final rules). Returns each rule record with its status label, the number of systems and the ECHO population served with at least one UCMR 5 sampling-location average at or above a scenario limit's comparison cutoff (EPA Table 4 method), the PWSIDs whose only such averages are against limits proposed for rescission, the largest matching systems by population up to limit, the register's published figures for comparison and the method version, optionally filtered by state file, owner type and minimum population. These are technical comparisons of monitoring averages, not compliance results (not_compliance: true).
pfas_system_history
ChatGPTReturns one public water system's full record in the MORVS PFAS Drinking-Water Register by PWSID: each sampling location (a hashed key) with its sample events, first and last dates and complete-set averages in ppt, per-chemical sample and detection counts with the highest single result, the Hazard Index, the six derived flags, the six 2024 rule records with each one's status history and whether this system's averages meet its comparison cutoff, the source file hashes and the register's caveats. Raw samples are not published. UCMR 5 results are entry-point monitoring, not compliance determinations or household tap results, and EPA's May 2026 rescission and deadline documents are proposals. A system whose name the register withholds is shown by PWSID and city.
pfas_system_lookup
ChatGPTLooks up public water systems in the MORVS PFAS Drinking-Water Register, built from EPA's final UCMR 5 monitoring file, the ECHO SDWA system file and the Federal Register. Give exactly one of pwsid (one system), zip (every system that reported serving that ZIP code; a service-area hint, not a confirmed supplier) or state (one state file, or TRIBAL), optionally with name to filter a state's systems by system name or city. Each match carries the PWSID, display name, state, city, ECHO population served, owner type and its label, activity status, the six derived flags, the 2024 limits whose comparison cutoff at least one sampling-location average meets, and for ZIP queries the match scope; the answer carries source dates, the status of each limit named and the register's caveats. UCMR 5 results are entry-point monitoring samples compared with the 2024 limits by EPA's Table 4 method, not compliance determinations or household tap results, and EPA's May 2026 rescission and deadline documents are proposals. A system whose name the register withholds is shown by PWSID and city.
pharmacy_340b_clients
ChatGPTReverse 340B lookup: given a contract-pharmacy ID, list every 340B covered entity that dispenses its discounted drugs through that pharmacy — reveals how concentrated the contract-pharmacy network is around big chains. Pharmacy IDs come from a covered entity's detail page.
plate_reader_decisions
ChatGPTFinds official city and county automated license plate reader matters in the MORVS License Plate Reader Register. Returns selected published decision fields with complete recorded action timelines, source links, vendors named in titles, and dollar figures exactly as printed. Filter by state, jurisdiction, vendor, recorded motion outcome, subject, decisive-action year or title substring query. Outcome means only the motion as recorded. An approval is an authorization, not proof of a signed contract or installed camera. A failed motion does not mean a contract ended. Amounts are printed figures, often ceilings, never money spent. No row means no qualifying action was found in the checked sources and period; it is not proof that no ALPR use exists.
plate_reader_state_law
ChatGPTReturns the MORVS License Plate Reader Register's full state-law research record for one US state or DC, including official law citations and URLs, checked date, verification status, retention and sharing provisions, exceptions and search limitations. A "none_found" or unverified status describes the sources checked; it does not prove that no applicable law exists. Verify the current official text before relying on a provision.
police_grant_awards
ChatGPTSearch the Police Grant Register's DOJ COPS award lists by fiscal year (FY2020 to the latest published, default latest), program, state and recipient name. Amounts are DOJ's announced awards, not outlays; the lists do not record scoring, so an award must never be attributed to a recipient's immigration posture. DOJ's Estimated Award Amount columns are labelled estimated. The FY2024 technology rows are held because that list does not reconcile; DOJ's printed total is identified as its figure only. JAG formula allocations, COPS awards and forfeiture payments are never added together. Reviewed DOJ-list and ICE 287(g) matches are labelled as such; forfeiture payments appear only side by side.
police_grant_recipient
ChatGPTFind institutional recipients by name substring in the Police Grant Register's DOJ COPS award lists from FY2020 to the latest published fiscal year. Returns each recipient's award rows by program and fiscal year, reviewed DOJ-list and ICE 287(g) match labels, and any DOJ equitable-sharing payments side by side. COPS, JAG formula allocations and forfeiture payments are never added. 'No COPS award appears in the published lists' never means an agency gets no federal money; absence covers only the published COPS lists. Award lists do not record application scoring or money paid out.
prediction_market_exchange
ChatGPTFind a CFTC designated contract market (DCM) or pending application by exchange code or name substring in the MORVS Prediction Market Register. Returns CFTC-listed status/date and fixed-phrase remark flags, filing counts by listed year/category/status, stay and withdrawal counts, approval-pending and perpetual-futures counts, code-map evidence and open federal comment periods. Pending means only pending; a stay is a CFTC remark, not a finding. Filing counts measure filing practice, not markets, trading, users or odds. Dates are as listed by the CFTC and may be re-dated.
prediction_market_filings
ChatGPTSearch event-type CFTC product filings in the MORVS Prediction Market Register by exchange code/name, listed year, filing status, MORVS topic, CFTC stay remark or public title substring. The default year is the latest published listed-year shard. A Certified filing was self-certified by the exchange under CFTC rule 40.2; it was not approved by the CFTC. Approval Pending and Approved are a separate product-approval route. Filings are not markets, volume, users or odds. Dates are as listed by the CFTC and can be re-dated after a status change. Topics are MORVS classification. Titles that may name individuals are withheld and are never searched; use the CFTC ID link for the source record. A stay is not a finding of a violation.
resolve
ChatGPTEntity locator: resolve a company name/ticker/CIK to its canonical SEC identity + aliases, and an available_in map of which federal datasets contain it (with links). Use first to find an entity and discover where it has data.
sanctions_check
ChatGPTSearch MORVS's partial local OFAC SDN index by case-insensitive substring, including indexed aliases. Returns hit=true for a local match. A non-match is not sanctions clearance; check /coverage and OFAC's current official lists.
sanctuary_doj_list
ChatGPTSearch DOJ’s current list page or its August 2025 announcement by name; defaults to the current page. “Sanctuary jurisdictions” is DOJ’s label, not a finding by MORVS. The two DOJ pages differ; removed entries carry DOJ’s linked release and stated reason. A county-level 287(g) agreement and a state-level designation can both be true; no conclusion is drawn from the two records. A non-match says only that a name was not found on the selected DOJ page; DOJ says its list is not exhaustive.
sanctuary_state_reports
ChatGPTSearch published California Values Act transfer counts, Illinois Way Forward Act detainer, warrant, transfer and request reports, or Oregon Sanctuary Promise Act requests and communications. These are agency self-reports. State measures differ and are never added or ranked across states. State is required; agency name, year and result limit are optional. Oregon’s year is the end year of its June–May reporting period. Illinois agency rows are available for 2024; 2023 has statewide figures and a separate non-filer list, not agency appendix rows. A zero is a reported zero, a null is absent from an appendix. “No transfers reported” and “did not file” do not mean “no cooperation”. Never say an agency “violated” a law or acted “illegal” based on these records.
screen_all
ChatGPTExperimental lookup of one name across MORVS's local public-record indexes; availability is not guaranteed. Returns per-dataset match counts and an indicative 0-100 score (risk_score) summarizing how many indexed records matched. The score is not a risk rating or compliance determination, and a low score or non-match is not clearance. Coverage and freshness vary by dataset.
screen_batch
ChatGPTScreen up to 50 entities in one call. Returns a per-entity risk_score + risk_level + per-vertical summary for each, plus the list sorted highest-risk-first. Use to vet a whole vendor / counterparty / portfolio list at once.
search
ChatGPTFull-text search across all parsed SEC/FDA/register/weather section bodies (BM25 ranking). Also resolves company names for ticker lookup.
search_340b
ChatGPTSearch HRSA 340B Drug Pricing Program covered entities (disproportionate-share hospitals, FQHCs, clinics buying federally-discounted outpatient drugs) by name. Returns 340B ID, entity type, state, participating status, and contract-pharmacy count. ~93k entities.
self_driving_company
ChatGPTFind reporting companies in the MORVS Self-Driving Crash Register by published name or slug. Returns separate ADS and ADAS company profiles, report and incident counts, severity, location, confidentiality index, investigations, recalls and California DMV permits. share_of_kind is a share of reports, not a crash rate. A DMV permit holder may be a different legal entity; the register notes that difference. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
self_driving_reports
ChatGPTSearch NHTSA Standing General Order crash-report rows in the MORVS Self-Driving Crash Register. Requires ADS or Level 2 ADAS and can filter by company, state, alleged injury severity, crash month and the year NHTSA first received the report. Returns latest report versions with NHTSA CSV citations, not narrative text. No report has its own NHTSA web page. OTHER filings are never returned as crashes. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
short_interest_search
ChatGPTLook up a ticker's FINRA short interest: shares sold short, days-to-cover, change vs prior settlement. The real short metric (distinct from short volume which includes market-maker hedging). Pairs with fails-to-deliver + short volume. FINRA source terms: https://www.finra.org/terms-of-use.
short_interest_squeeze
ChatGPTFINRA equity short interest: securities with the highest days-to-cover (the classic short-squeeze indicator) + the largest absolute short positions, from the latest bi-monthly settlement. TRUE short interest (not short volume). Use to find squeeze candidates and the most-shorted stocks. FINRA source terms: https://www.finra.org/terms-of-use.
short_squeeze_candidates
ChatGPTTop short-squeeze candidates by a MORVS (AI Analytics LLC) 0-100 pressure score combining FINRA days-to-cover + short-sale volume % + SEC fails-to-deliver per symbol. The methodology is separate from and does not relicense FINRA inputs. Transparent heuristic, not investment advice. FINRA source terms: https://www.finra.org/terms-of-use. Optional min_score filter.
short_volume_most_shorted
ChatGPTFINRA short-sale volume: the most-shorted liquid stocks by average short % of volume (rolling ~60 trading days), plus dataset summary. Pairs with fails-to-deliver for the short-selling picture. Caveat: FINRA short volume includes market-maker hedging, not short interest. FINRA source terms: https://www.finra.org/terms-of-use.
short_volume_search
ChatGPTLook up a ticker's FINRA short-sale volume: average / latest / peak short % of consolidated volume over the rolling ~60-trading-day window. FINRA source terms: https://www.finra.org/terms-of-use.
super_pac_donors_donor
ChatGPTLook up an organisation that gave to super PACs or hybrid PACs in the Super PAC Donors Register by name substring, matched against the reviewed name, every name as filed and the reviewed company family, optionally for one FEC cycle (2026 for 2025-26, or 2024 for 2023-24; both when omitted, never added across cycles). Returns up to 5 named organisations, each with its reviewed class and status, its company family and sector, and the parent or group record from an SEC filing where the register cites one; and by cycle, its contributions by recipient super PAC as the FEC lists it, with the committee id, refunds kept apart and never netted, dates as filed and the FEC bulk file date. Beside them, never added or put in sequence: for the company families in the register's side-by-side table, the connected PAC's contributions to federal candidates' committees with its all-committees summary line (FEC), contract obligations by fiscal year (USAspending) and LDA filing counts; the organisation's net donation to the 2025 inaugural committee where it is on both records (Form 13); and ICE and CBP contract obligations for Pay-to-Play Register contractors. An empty column is a search result, not a fact about the organisation. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. A name that matches only a withheld or unreviewed string answers exactly like a name that matches nothing. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
super_pac_donors_recipient
ChatGPTLook up a super PAC or hybrid PAC in the Super PAC Donors Register by committee name substring, as the FEC committee master lists it, or by its committee id (C followed by 8 digits). Returns up to 5 committees, each for the 2025-26 and 2023-24 cycles with its name, id and committee type as the FEC lists them, and three measures, each labelled and never combined into one figure: organisation contributions by reviewed class (issue and social-welfare groups split by IRS match; withheld and unreviewed money as amounts only), contributions from committees, and itemized contributions from individuals, counted only; conduit pass-through rows (ActBlue, WinRed) are shown apart under their own label and are excluded from every total. Each cycle lists the largest named organisation donors to that committee and the committee's own FEC summary line. Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. Organisation, committee and individual money into super PACs are three measures, each labelled and never combined into one figure. Organisation money is not corporate money: every total states its class. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
super_pac_donors_summary
ChatGPTTotals from the Super PAC Donors Register, built from the FEC's bulk files: organisation contributions to super PACs and hybrid PACs for the 2025-26 and 2023-24 cycles, with rows, names as filed, named donors, the reviewed share and each FEC bulk file's date, and the change over the same date window in both cycles; organisation money by reviewed class, with issue and social-welfare groups split between those an exact IRS Exempt Organizations record matches and those whose status was not matched, beside companies, unions, trade associations, committees, tribal governments and other organisations; withheld and unreviewed strings as counts and amounts only; itemized contributions from individuals reported directly to these committees, counted only, beside organisation money on that exact basis, with conduit pass-through rows (ActBlue, WinRed) shown apart and excluded from every total; sectors and the largest recipient committees as the register lists them; the lag line; and the register's headlines with their basis. Organisation, committee and individual money into super PACs are three measures, each labelled and never combined into one figure. Organisation money is not corporate money: every total states its class. Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
trade_remedy_search
ChatGPTSearch USITC trade-remedy investigations — antidumping/countervailing-duty (AD/CVD) + Section 337 unfair-import (patent/IP) cases — by investigation number, product, or title. ~4,386 cases; the trade-enforcement vertical (companies/countries fighting over dumped, subsidized, or IP-infringing imports).
ukraine_money_recipient
ChatGPTFind institutions and companies in the Ukraine Money Register by name substring: named award-record recipients of Ukraine-division accounts (obligated by fund code, obligated and outlaid by class, place of performance) and DoD contracts whose own records mention Ukraine, grouped by parent as USAspending lists it. With include_loop (default true), loop companies show three records side by side, never added and with no causation: DoD Ukraine-mention contract obligations; connected-PAC contributions to federal candidate committees by cycle, with the top recipient committees named exactly as the FEC lists them, with party and office; and lobbying filings citing Ukraine bills, each with its all-issue amount. Corporate PAC money is voluntary contributions from people the PAC may solicit, not company treasury or contract money; the FEC's own explanations are quoted. No individual donors, and no withheld or individual award recipients. The published recipient list holds the largest named institutions only, so a non-match is not evidence that no award was made.
ukraine_money_totals
ChatGPTTotals from the Ukraine Money Register of federal Ukraine-response money. view=stages (default): the inspectors general's made available, obligated and disbursed (outlaid) totals as printed, each with its own as-of date; acts: the five supplemental acts' made-available totals re-added clause by clause from the act text, beside SIG-OAR's table; codes: USAspending's emergency fund codes 3, 6, AAA, AAB and AAJ, where the mixed codes are split against the act text into Ukraine-only, probable and shared accounts, shown apart; places: place of performance in the award records; oversight: inspector-general and GAO findings quoted exactly with report numbers, dates and pages. Stages (made available, obligated, outlaid, announced) are never added to one another. Pentagon contracts whose own records mention Ukraine are not tied to the supplemental appropriations, and USAspending links only $0.47 billion of DoD's $121.30 billion obligated under the five codes to named awards. Words alleging loss or wrongdoing, such as fraud or diversion, appear only inside exact quotes.
upcoming_earnings
ChatGPTEstimates each company's next 10-Q/10-K filing date from its past SEC filing cadence. Dates are estimates, not announced filing dates.
upcoming_sales_pressure
ChatGPTLists companies by the size of insider sales planned in recent SEC Form 144 notices (filed intentions to sell, not completed trades).
watchlist
ChatGPTConsolidated activity rollup for a list of tickers — insider 30d net flow, Form 144 pressure, recent material events, cluster-sell flag. Up to 25 tickers per call.
what_your_pension_owns
ChatGPTWhat public pension funds reported holding in companies on US government lists or in MORVS's registers, from What Your Pension Owns, a MORVS register built from the SEC's quarterly Form 13F data sets (Q3 2025 to Q2 2026 at launch) and five funds' own holdings reports, matched issuer by reviewed issuer to the Pentagon's Section 1260H list, Treasury's NS-CMIC list, DHS's UFLPA Entity List and two MORVS registers (the Detention Ledger's publicly traded operators, and the listed parents of vendors counted in Who Gets Your Data's government lane). Optional filters: fund (a reviewed fund's name, a two-letter state or a 10-digit SEC CIK), issuer (a company name, a ticker the register carries, or a 9-character CUSIP), list (a list ID the register carries, or its title), source (13f or report) and quarter (a quarter-end date the register carries, YYYY-MM-DD). With filters: for each list, the matched holdings by source, date and security kind, Form 13F rows and the funds' own report rows kept apart and debt apart from shares, each group's value split into holdings on the list at the holding date and holdings designated later. When the matches are one fund's: up to 25 of its positions, the two sources kept apart, latest holding date leading (holding date, issuer, shares or debt, the value as filed or printed, each list's names as printed with its designation or register date and the designation-after-holding flag, the register's claim sentence, and the filing or report rows behind it). When they span funds: each fund with its count of positions and the arguments that return its positions (for_positions). With no filters: the register's headlines with their basis, each list's latest-quarter Form 13F figures, and the counter-findings (how many funds' filings hold no reviewed match to each list). The two sources are never added together, and no total adds two lists, two dates, two reports, or shares and debt; a company on two lists appears under each. A holding is a position a fund reported on a stated date: it does not show why the fund held it, whether it followed an index, or what the fund holds today. A designation dated after a holding is flagged on that position (designation_postdates_holding) and its value is kept apart from holdings on the list at the holding date. The MORVS registers are not government designations. No state divestment law is applied. Answers are kept under about 25 KB. The answer carries as_of, retrieved, the file dates, the question, list, index, coverage and law notes, the claim rules, the licence and the source page on morvs.ai; an empty result carries the register's absence note (no reviewed match in that fund's filing or report for that date, which is not a statement about other dates, other accounts or holdings through outside funds). Pension funds are institutions; no person is named.
what_your_pension_owns_fund
ChatGPTOne reviewed public pension fund in What Your Pension Owns, a MORVS register built from the SEC's quarterly Form 13F data sets (Q3 2025 to Q2 2026 at launch) and five funds' own holdings reports, matched issuer by reviewed issuer to the Pentagon's Section 1260H list, Treasury's NS-CMIC list, DHS's UFLPA Entity List and two MORVS registers (the Detention Ledger's publicly traded operators, and the listed parents of vendors counted in Who Gets Your Data's government lane), found by fund (its name as the register prints it or as filed, a two-letter state, or a 10-digit SEC CIK; when several match, each in short form with the arguments that return its full profile). Returns the fund's profile (state, type and its label, the basis for counting it as a public pension fund, CIK, legal name as filed and SEC filer link, and its own holdings report where the register reads one, with its as-of date, unit and scope), its Form 13F filing for each quarter (accession, amendments, table entries and value as filed), its figures by list, source and date and up to 25 matched positions (shares and debt apart, with each list's date and the designation-after-holding flag), the register's quarter-to-quarter changes for the fund (positions added and exited per list between quarters it filed, with the values on each date side by side, never subtracted), and, for each list where it holds no reviewed match, the counter-finding with the quarters and report read. The two sources are never added together, and no total adds two lists, two dates, two reports, or shares and debt; a company on two lists appears under each. A holding is a position a fund reported on a stated date: it does not show why the fund held it, whether it followed an index, or what the fund holds today. A designation dated after a holding is flagged on that position (designation_postdates_holding) and its value is kept apart from holdings on the list at the holding date. The MORVS registers are not government designations. No state divestment law is applied. Answers are kept under about 25 KB. The answer carries as_of, retrieved, the file dates, the question, list, index, coverage and law notes, the claim rules, the licence and the source page on morvs.ai; an empty result carries the register's absence note (no reviewed match in that fund's filing or report for that date, which is not a statement about other dates, other accounts or holdings through outside funds). Pension funds are institutions; no person is named.
who_gets_your_data_government
ChatGPTFederal purchases in the "Bought by the government" lane of Who Gets Your Data, a MORVS pilot register built from USAspending prime contract transactions (award types A to D) by action-date fiscal year over the lane's window (FY2017 to FY2026 at launch): what agencies obligated to the reviewed vendors and resellers of the lane's capability classes (nine at launch; ids as the register carries them, such as device_forensics, people_search, comms_analysis, social_media_osint, plate_data, face_recognition, cell_site_simulators and location_data). Optional filters: agency (an agency or sub-agency name, or an abbreviation derived from the names, such as ICE or FBI), vendor (a recipient as filed, its parent, a reviewed vendor family or a recipient UEI, such as Cellebrite or LexisNexis), capability (a class id or its label, as the register carries them) and fiscal_year (an action-date fiscal year the register carries). With filters: the matching counted totals by fiscal year and by capability, bundles kept beside, the count of matching awards awaiting review, and up to 25 awards, largest matching obligation leading (PIID, agency, recipient as filed, fiscal years, the amount obligated, each matching transaction's capability and basis, the matched words and the words as published or "withheld", the outlay for the whole award where the record gives it, and the USAspending link). With no filters: the lane's headlines with their basis, the totals by fiscal year and by capability, and the top buyers. Amounts are obligations net of de-obligations, never outlays: an outlay appears only where the award record gives one, labelled as the outlay for the whole award. A contract is not proof of use on any person. Capability comes from the award's own words or a single-product vendor's product line, each with its basis (transaction_words, award_base_words, vendor_product or review); a bundle naming two capabilities sits in the lane total and in no capability total. The floor year (FY2026 at launch) is a floor, as Defense Department actions reach FPDS late, and each figure for it is flagged. A reviewed universe, not a complete inventory. An empty result returns the lane's absence note: no award in this reviewed universe, which says nothing about purchases outside it. The lane is a separate table: never added to the benefits and identity panel, the context (Palantir, Dataminr Defense Department licences), the excluded amounts or the handoff, broker and legal-duty tables. The answer carries as_of, the file dates, the source page on morvs.ai, the licence and the lane's universe, obligation, use, capability, floor, absence, context and panel notes. Recipients as filed, organisations only: the register drops sole proprietors and person-shaped recipients, and award words that could name a person are published as "withheld".
who_gets_your_data_government_award
ChatGPTOne federal award in the "Bought by the government" lane of Who Gets Your Data, a MORVS pilot register built from USAspending prime contract transactions (award types A to D) by action-date fiscal year over the lane's window (FY2017 to FY2026 at launch), found by its PIID (6 to 40 letters, digits or hyphens, case ignored) through the register's lookup file. Returns each transaction with its date, fiscal year, amount obligated, label (a capability, a bundle, or a label that is never counted), basis, the matched words and the words as published or "withheld"; the award's review (label, verified words and note); its net obligations by fiscal year; and the outlay for the whole award where the award record gives it. It says whether the award is counted in the lane, awaiting review (never counted until reviewed), not counted (words mixed with another product), unclassified or in the separate benefits and identity panel. A PIID can name more than one award; each is returned. Amounts are obligations net of de-obligations, never outlays: an outlay appears only where the award record gives one, labelled as the outlay for the whole award. A contract is not proof of use on any person. Capability comes from the award's own words or a single-product vendor's product line, each with its basis (transaction_words, award_base_words, vendor_product or review); a bundle naming two capabilities sits in the lane total and in no capability total. The floor year (FY2026 at launch) is a floor, as Defense Department actions reach FPDS late, and each figure for it is flagged. A reviewed universe, not a complete inventory. An empty result returns the lane's absence note: no award in this reviewed universe, which says nothing about purchases outside it. The lane is a separate table: never added to the benefits and identity panel, the context (Palantir, Dataminr Defense Department licences), the excluded amounts or the handoff, broker and legal-duty tables. The answer carries as_of, the file dates, the source page on morvs.ai, the licence and the lane's universe, obligation, use, capability, floor, absence, context and panel notes. Recipients as filed, organisations only: the register drops sole proprietors and person-shaped recipients, and award words that could name a person are published as "withheld".
who_gets_your_data_handoffs
ChatGPTList documented data handoffs in Who Gets Your Data, a MORVS pilot register: edges from a source organisation and product, through a data type, to a recipient, each from one primary document (an FTC complaint or order, a Federal Register notice, a court opinion or order). Optional filters: data_type (a substring of the data type in the document's words), legal_status (a status key the register carries, such as allegation_in_complaint, finding, order_terms or request_count; matched against the row's own status and its settlement and remedy statuses, with the match shown), recipient (a substring of the recipient as worded or of a named recipient organisation) and year (the primary document's year, a year the register carries). Up to 25 edges in published order, with the full count and the counts by the rows' own legal status. Each handoff gives the source organisations, product and brands, the data type in the document's words, the recipient exactly as worded (named, or the document's own recipient class), the conduct period as worded, the verbatim quote, the document's type, issuer, date and URL, and a settlement, remedy, current status or opt-out only where a document states one. Each handoff carries its primary document's own legal status with the register's label exactly as published (a court's finding, an allegation in a complaint, a settlement with or without an admission, an order's terms, or a count of requests): an allegation is not a finding, a request is not a disclosure, and old conduct is not current practice. An empty result returns the register's absence note: no documented record in this corpus, which says nothing about what a product or company does now. The three tables (handoffs, broker registry answers, legal duties) hold different kinds of evidence and are never added together, scored or ranked; no answer gives a total across them. A pilot: a small, hand-checked set of documented handoffs, the brokers' own registry answers and one legal-duty exhibit, in three separate tables; it is not a census of data sharing. The answer carries the pilot note, lane note, coverage line, as-of date, file dates, legal-status labels, the register's rules, its licence and the source page on morvs.ai. Organisations, companies, agencies and governments only: no individual is named, and there is no bank or personal-report lookup.
who_gets_your_data_lookup
ChatGPTLook up a brand, product, company or recipient, or a 5-digit ZIP code, in Who Gets Your Data, a MORVS pilot register built from FTC complaints and orders, Federal Register notices, a Supreme Court opinion, a Sacramento court order, FinCEN's Year in Review and one Geographic Targeting Order, and the California data broker registry as MORVS's Data Broker Register republishes it. Names are matched against the register's published lookup keys (case, accents, punctuation and legal-form words such as Inc. or LLC ignored); a ZIP is matched against the order's list. Returns the matching rows grouped by table, each table with its own coverage line and notes: handoffs (up to 25 edges with their full fields), broker registry answers (up to 25 brokers, each matched registration with the questions it answered yes), legal duties (a ZIP's covered-location row with the order's dates and its terms and exceptions as quoted), and "Bought by the government" (agency, sub-agency, vendor family, recipient, reseller, capability, fiscal year as FY2026 or award PIID keys, and agency abbreviations such as ICE derived from the names: each row with the lane's published obligations by fiscal year, the floor year flagged, or the award's standing), with benefits and identity panel matches in their own group. Each handoff carries its primary document's own legal status with the register's label exactly as published (a court's finding, an allegation in a complaint, a settlement with or without an admission, an order's terms, or a count of requests): an allegation is not a finding, a request is not a disclosure, and old conduct is not current practice. Each handoff gives the source organisations, product and brands, the data type in the document's words, the recipient exactly as worded (named, or the document's own recipient class), the conduct period as worded, the verbatim quote, the document's type, issuer, date and URL, and a settlement, remedy, current status or opt-out only where a document states one. An empty result returns the register's absence note: no documented record in this corpus, which says nothing about what a product or company does now. A ZIP result is a covered business location under one FinCEN Geographic Targeting Order, subject to its exceptions, and carries the register's ZIP note: it says nothing about any resident, account or transaction. Broker rows are each business's own answer to the California registry, with the question as the form words it, and carry the broker note: no broker is accused of anything. The three tables (handoffs, broker registry answers, legal duties) hold different kinds of evidence and are never added together, scored or ranked; no answer gives a total across them. The lane is a separate table: never added to the benefits and identity panel, the context (Palantir, Dataminr Defense Department licences), the excluded amounts or the handoff, broker and legal-duty tables. A pilot: a small, hand-checked set of documented handoffs, the brokers' own registry answers and one legal-duty exhibit, in three separate tables; it is not a census of data sharing. The answer carries the pilot note, lane note, coverage line, as-of date, file dates, legal-status labels, the register's rules, its licence and the source page on morvs.ai. Organisations, companies, agencies and governments only: no individual is named, and there is no bank or personal-report lookup.
who_gets_your_data_summary
ChatGPTFigures from Who Gets Your Data, a MORVS pilot register built from FTC complaints and orders, Federal Register notices, a Supreme Court opinion, a Sacramento court order, FinCEN's Year in Review and one Geographic Targeting Order, and the California data broker registry as MORVS's Data Broker Register republishes it, table by table and never added across tables: handoff edges and primary documents, counts by legal status with the register's labels, named and class-only recipients, rows with a settlement or a current status, and the review queue's counts (a candidate is never an edge until reviewed); the brokers' California registry answers, with each sharing question as the form words it, the registrations and distinct brokers answering yes, and the government or law-enforcement overlap; FinCEN's fiscal-year Bank Secrecy Act figures with their wording, page and unit, and the Geographic Targeting Order's dates, ZIP count by county group and quoted terms; the register's headlines with their basis, legal status and sources; what is not covered (documents left for manual download, hosts never fetched, the incumbents credited and the register's stated gap), with the stop rules; and, kept separate, the "Bought by the government" lane (obligations by fiscal year with the floor year flagged, by capability with basis, bundles, the top buyers, the people, location and plate-data group, awards pending review, the lane's headlines with their basis, and the mixed, unclassified, context, excluded and dropped amounts kept apart) and the benefits and identity panel with its own headline. FinCEN's figures count searches, reports and filers, never people, and carry the register's count note. Broker rows are each business's own answer to the California registry, with the question as the form words it, and carry the broker note: no broker is accused of anything. Each handoff carries its primary document's own legal status with the register's label exactly as published (a court's finding, an allegation in a complaint, a settlement with or without an admission, an order's terms, or a count of requests): an allegation is not a finding, a request is not a disclosure, and old conduct is not current practice. The three tables (handoffs, broker registry answers, legal duties) hold different kinds of evidence and are never added together, scored or ranked; no answer gives a total across them. The lane is a separate table: never added to the benefits and identity panel, the context (Palantir, Dataminr Defense Department licences), the excluded amounts or the handoff, broker and legal-duty tables. Amounts are obligations net of de-obligations, never outlays: an outlay appears only where the award record gives one, labelled as the outlay for the whole award. A contract is not proof of use on any person. Capability comes from the award's own words or a single-product vendor's product line, each with its basis (transaction_words, award_base_words, vendor_product or review); a bundle naming two capabilities sits in the lane total and in no capability total. The floor year (FY2026 at launch) is a floor, as Defense Department actions reach FPDS late, and each figure for it is flagged. A reviewed universe, not a complete inventory. A pilot: a small, hand-checked set of documented handoffs, the brokers' own registry answers and one legal-duty exhibit, in three separate tables; it is not a census of data sharing. The answer carries the pilot note, lane note, coverage line, as-of date, file dates, legal-status labels, the register's rules, its licence and the source page on morvs.ai. Organisations, companies, agencies and governments only: no individual is named, and there is no bank or personal-report lookup.
wiretap_orders
ChatGPTFinds individual court-authorized wiretap orders as published in one annual US Courts Wiretap Report. Defaults to the latest report year; filters use the report year, court, state, jurisdiction, offense group and intercept type. A row was reported to the courts in that report year and may have ended earlier. Results and costs are per order as reported. The annual headline count undercounts that year because some orders are reported in later years. Related orders can share people, intercepts and arrests, so do not sum those fields across rows. Zero or blank means none reported, not that nothing happened. Flags identify missing prosecutor reports and implausible figures. A missing row is not proof that no surveillance occurred; the published order shards start in 2016.
adviser_search
ClaudeSearch SEC-registered investment advisers (RIAs + Exempt Reporting Advisers) by firm name. Returns CRD number, type, regulatory assets under management, state, and whether the firm has Form ADV Item 11 disciplinary disclosures. ~23.4k advisers covered.
advisers_flagged
ClaudeSEC-registered investment advisers reporting Form ADV Item 11 disciplinary disclosures (criminal / regulatory / civil), ranked by disclosure count. ~2,291 advisers flagged. Filter by state, type (RIA/ERA), or kind (criminal/regulatory/civil).
ag_census_county_search
ClaudeSearch USDA Census of Agriculture by county/state name. Returns the place_id ({state_fips}-{county_code}) to pull a county's farm economy — number of farms, land in farms, market value of products sold, net farm income, and top commodities by sales (2022). Aggregate federal statistic — zero PII.
border_force_area
ClaudeLook up a CBP area of responsibility in the Border Force Register, built from CBP's own monthly use-of-force and assault files (FY2021 on), by sector or field-office name substring. An area is a CBP office's area or unit (a Border Patrol sector, an OFO field office or a national unit), never a city or an operation; CBP's own service-area statement is quoted where CBP gives one. Returns yearly figures, or monthly figures for one fiscal_year: use-of-force incidents with their force types (force totals include vehicle and vessel force, such as tire-deflation devices), officers/agents using force, assault incidents, and officers/agents assaulted. Officers assaulted and officers using force are counted per incident: one officer in two incidents counts twice, and neither is a count of people injured. The current fiscal year is labelled partial. No rates (nothing is divided by encounters or staffing) and no causation: CBP's files do not say why counts change. A non-match says only that the name is not in CBP's files.
border_force_national
ClaudeNational yearly totals from CBP's own use-of-force and assault files in the Border Force Register (FY2021 on; the current fiscal year is labelled partial): use-of-force incidents by force type (vehicle and vessel force included), officers/agents using force and officers/agents assaulted (both counted per incident), and assault incidents, each year with the CBP file and date it comes from. With fiscal_year, adds that year's revision log entries, each labelled as published on its date: revisions are as-published versions of CBP's files, not corrections and not evidence of misconduct. Also returns the status of DHS's department-wide use-of-force tables, the series with no newer edition (including ICE's count in DHS's table), and DHS's tables beside CBP's files with the definitions note; the two use different definitions and are never added or merged. No rates and no causation.
breach_notice_organization
ClaudeFind public notifying organisations by name substring in the Washington, Delaware and California Attorney General breach-notice lists. Returns notice IDs and counts separately for each state, plus any paired Washington and Delaware notices whose notice dates are within three days. When paired discovery dates differ, the two states' records carry different dates; neither date is declared correct. Withheld names are excluded from this index and never returned or searchable. A non-match in these three states' records never means no breach: list coverage, reporting thresholds and organisation names vary. The notifying organisation may differ from the breached one.
breach_notices
ClaudeSearch the Washington, Delaware or California Attorney General breach-notice list by public notifying-organisation name and year (notice year for Washington and Delaware, reported year for California). Timing uses only the dates in each record: Washington and Delaware results label whether the interval is beyond that state's statutory window by the record's dates. The records do not show permitted delays, including law-enforcement requests, or any enforcement outcome; the timing flag is no legal finding. Never call a notice "late" or say an organisation "violated" the law. California has no discovery date, so no timing is computed. Resident counts are notices sent to residents, not unique people, and are never added across states; Delaware also reports the company's stated total affected separately. A notifying organisation may differ from the breached one. Names that may identify individuals are withheld, never returned or searchable. Absence from these three lists is no evidence that an organisation had no breach.
broadcast_station_search
ClaudeSearch the FCC broadcast station directory by call sign or community of license. ~109,812 US over-the-air facilities (TV/FM/AM/LPTV/translators, ~32,883 active). Returns service, channel, frequency, community, status, and license expiration. Station facts only — no ownership.
bsee_search
ClaudeSearch BSEE offshore oil & gas operators (US Outer Continental Shelf — Gulf of Mexico, Pacific, Alaska) by name. Returns each operator's incident-enforcement totals: warnings, component shut-ins, facility shut-ins. ~674 operators, 1948-2026.
bsee_worst_operators
ClaudeRanks offshore oil and gas operators by counts of BSEE incidents of noncompliance: facility shut-ins (default), component shut-ins, total events, or warnings. Counts come from BSEE records and are not adjusted for operator size or activity.
building_permits_search
ClaudeSearch Census Building Permits Survey by county/state name. Returns place_id ({state_fips}-{county_fips}) to pull a county's residential-construction permit history — buildings, housing units, and valuation by structure type (1-unit / 2-units / 3-4 / 5+), 2015-2025. Aggregate federal statistic — zero PII.
cfpb_by_company
ClaudeSearch MORVS's partial local CFPB complaint index by company. Returns indexed complaint counts, products, and narratives only where the CFPB published them. Check /coverage; this is not the full CFPB database.
chemical_accident_search
ClaudeSearch EPA RMP hazardous-chemical facilities (Clean Air Act 112(r)) by facility or parent-company name. Returns each facility's chemical-accident count + location. ~99k facilities — the chemical-plant accident ledger EPA keeps FOIA-only (releases, fires, explosions, deaths).
chemical_accident_worst
ClaudeRanks parent companies by harm reported in EPA Risk Management Program accident histories: deaths (default), injuries, or accident count. Counts come from RMP filings and are not adjusted for company size or number of facilities.
china_ai_filings
ClaudeSearch the CAC public generative AI, deep-synthesis and algorithm lists by a Chinese or English substring in the listed service or algorithm name, filer legal name, curated filer English name, or CAC filing number. Registry defaults to all; province applies to generative AI rows with a CAC province field. A us_list=true filter uses confirmed tier A legal-entity crosswalks only; brand uses an exact filer legal-name allowlist. Returns the CAC number, name and filer as printed, a separately labelled own English name or MORVS translation when available, CAC notice, date and its date type. The CAC says completed filing (备案) or registered (登记); neither grants approval. Filings are neither distinct models nor distinct companies. A brand in a filer legal name does not establish ownership. No match does not establish that a service or company is absent from China.
china_ai_us_lists
ClaudeRead the curated crosswalk from CAC filer legal names to the US Entity List and Treasury NS-CMIC list. Search by Chinese or English entity-name substring; tier defaults to A. Tier A has published evidence for a legal-entity match. Tier B is a possible match, not confirmed, and is outside headline counts. Each entry gives the list name, start date, evidence type and text, and matching CAC record IDs. The Entity List imposes export-licence requirements for specified items subject to the EAR; the NS-CMIC investment ban covers specified US-person purchases or sales of publicly traded securities. These are different restrictions, not generic sanctions. The separate US actions timeline includes the DHS DeepSeek order with Federal Register links. MORVS does not determine whether any CAC DeepSeek filer is the company named by DHS. Not being on either US list never means an entity is cleared.
clinical_trials_recent
ClaudeRecent clinical trial activity from ClinicalTrials.gov. Filter by phase, status, and sponsor. Use for drug pipeline questions.
cluster_buys
ClaudeCompanies where multiple distinct insiders bought open-market stock in the window. Strong conviction signal.
cluster_sells
ClaudeCompanies where multiple distinct insiders sold open-market stock in the window. Mass-exit detector.
common_owner_rollup
ClaudeTrace one organization owner across every Medicare facility it controls (hospital/hospice/home-health/skilled-nursing): facility_count, per-sector breakdown, PE/REIT flags, state spread. Takes a CMS owner_associate_id (from /api/v1/owners/largest or /owners/private-equity).
company_insider_pulse
ClaudeUnified Form 4 (executed) + Form 144 (proposed-sale) insider timeline for one company.
company_profile
ClaudeComprehensive dossier for one SEC-filing company: basic info, recent filings, insider activity windows (7/30/90d), upcoming Form 144 sales, latest financials (FY+quarterly), top 13F institutional holders, top mutual-fund holders, peers.
company_risk_flags
ClaudeMulti-source red-flag detector for one company — restatements (8-K item 4.02), auditor changes, officer departures, insider cluster sells, upcoming Form 144 pressure, Schedule 13D filings, FDA recalls.
compare_entities
ClaudeSide-by-side compliance-risk comparison of two entities. Runs the full 40-source screen for each and returns both risk scores, an overall verdict (who is riskier), the risk gap, and the biggest per-vertical differences. Answers 'which counterparty is riskier' in one call.
courts_by_party
ClaudeLitigation dossier for one company: docket count, top Natures of Suit, recent cases. Answers 'is this company being sued?' and 'for what?'
courts_recent
ClaudeRecent federal court dockets from CourtListener. Filter by party, court, or Nature of Suit (850 securities, 830 patent, 410 antitrust, 365/367 product liability, 480 consumer credit).
crop_insurance_search
ClaudeSearch USDA RMA crop-insurance Summary of Business by county/state name. Returns the place_id ({state_fips}-{county_fips}) to pull a county's underwriting history — liability, premium, subsidy, indemnity, loss ratio, top crops (1989-2026). Aggregate federal program statistic — zero PII.
cross_vertical_active
ClaudeDiscovery feed: list of entities with activity in 2+ regulatory verticals (SEC + FDA + contracts) inside the lookback window. Use to find entities worth pulling a full cross-vertical timeline for.
csb_investigation_search
ClaudeSearch U.S. Chemical Safety Board (CSB) investigations of major chemical incidents — explosions, toxic releases, fires — by company, location, or keyword. ~139 independent root-cause investigations (BP Texas City, West Fertilizer, DuPont, etc.); the deep-investigation layer over EPA RMP chemical accidents.
data_broker_check
ClaudeChecks whether a business is registered as a data broker in California, using the MORVS California Data Broker Register: the 2026, 2025 and 2024 registries filed with the California Privacy Protection Agency and the Attorney General's 2020-2023 list, keyed by website domain so a name change at the same domain reads as one history. Give a business name, a website domain or URL, or both; matching is by registrable domain or by normalised legal name, DBA or earlier filed name (case, punctuation and legal-suffix insensitive, exact after normalisation), and when several brokers match they are returned as separate candidates, never merged. Each match carries the broker's filings as filed: what it says it collects, whom it reports having sold or shared data to, the laws it says regulate it, its own consumer-request counts, its name history and its broker page. Registration is self-reported and California-only, so a non-match is not a finding that a business is not a data broker, is exempt from registration, or has never registered. Businesses only; queries shaped like a personal name are not run.
data_broker_diff
ClaudeCompares two files of the MORVS California Data Broker Register (the 2026, 2025 or 2024 CPPA registry, or the Attorney General's 2020-2023 list) under the register's published method: a registration in the earlier file whose normalised legal name has no registration in the later file is an apparent leaver; if its website domain does have a later registration it is counted as re-registered at the same domain, with the later names, otherwise as not re-registered under either name or domain; a later registration whose name and domain are both absent earlier is net new. Returns the counts and the named registrations in each group, plus the register's own published figures for consecutive files. A registration absent from a later file is not an exit, a merger or a finding: registration is self-reported and California-only, and a business may have stopped brokering, merged, changed domain or simply not filed.
data_demand_compare
ClaudeShows the latest headline row reported by each publisher in the MORVS Government Data Demand Register, side by side with each figure's label. The figures are each publisher's own counts for its own period and scope. Companies count differently; never add or rank their figures across companies. A missing figure means not reported, not zero. National-security figures remain bands. Aggregate counts cannot show whether any person's data was requested. Optionally restrict to one sector.
data_demand_source
ClaudeReturns one publisher's figures for one report in the MORVS Government Data Demand Register. Figures are that publisher's own counts for its own period and scope. Companies count differently. Never add or rank figures across companies. A missing figure means not reported, not zero. National-security figures remain bands. These aggregate counts cannot show whether any person's data was requested. Defaults to the latest complete report; optional filters select exact fact fields.
dataset_freshness
ClaudePer-dataset freshness across the federal data hub: FRESH/AGING/STALE/UNKNOWN vs each source's expected cadence, with age in days and separate last-successful-source-observation and indexed-content watermarks. Each row states which timestamp drives status. Optional status filter (FRESH/AGING/STALE/UNKNOWN).
debarments_check
ClaudeScreen a name against US federal contractor debarments (SAM Exclusions, ~10k entries) AND HHS OIG healthcare provider exclusions (LEIE, ~83k entries). Returns hit=true if any match. Use this for KYC + counterparty due diligence.
documented_vessel_search
ClaudeSearch USCG documented vessels by name or official number. ~394,495 US-flag vessels (recreational, fishing, freight, towing, passenger). Returns service, flag, tonnage, length, build year, hull material, hailing port, and certificate-of-documentation status. Vessel/asset facts only — no owner names or addresses (zero PII).
doj_by_company
ClaudeDOJ press release search for one company. Surfaces FCPA, antitrust, healthcare fraud, environmental crimes, and any other DOJ enforcement actions mentioning the name. Body-text substring match across 266k+ press releases since 2009.
doj_recent
ClaudeRecent DOJ Office of Public Affairs press releases. Filter by free-text query or DOJ component (Antitrust Division, USAO - SDNY, Criminal Division, ENRD, Civil Rights, etc.).
dossier
ClaudeCross-topic dossier for one entity — SEC filings, 13F holdings, mutual fund holdings, FDA recalls as firm, federal register agency mentions, federal contracts as recipient.
drone_company_check
ClaudeChecks whether a company or agency is named in the MORVS Drone Register, which compiles US government drone approval, restriction and compliance records: the DCMA Blue UAS Cleared List and Framework, FCC UAS Conditional Approvals, FAA Remote ID Declarations of Compliance, and the drone-linked entries of the Section 1260H list, the BIS Entity List and OFAC's NS-CMIC list. Matches by name or alias (case, punctuation and legal-suffix insensitive; substrings allowed) and returns each match's designations with the government record that names it, attached reference records, and federal contract totals where loaded. A non-match is reported as not found in the register; it is not clearance or a compliance determination.
drone_product_lookup
ClaudeFinds drone platforms, components and software by model, product or category text across the DCMA Blue UAS Cleared List and Framework, FCC UAS Conditional Approvals and FAA Remote ID Declarations of Compliance in the MORVS Drone Register. Returns each matching designation with the company, list, action, date and source record, plus the current FCC Covered List exemption context taken from the register's reference records. An FAA declaration is the manufacturer's own statement, not an approval.
drug_label
ClaudeLook up an FDA drug label by brand, generic, or substance name. Returns indications, dosage, warnings, adverse reactions, drug interactions.
election_equipment_jurisdiction
ClaudeFind a jurisdiction by two-letter state code and literal jurisdiction name substring. Returns coded equipment answers self-reported in the EAC's 2024 Election Administration and Voting Survey (November 2024), with make/model and MORVS-curated vendor labels. Missing and coded Other answers do not mean no equipment. The survey does not establish equipment used on 2026-11-03; check Verified Voting and the state election office. Vendor counts overlap and must not be summed. Say voters participated in jurisdictions that reported a vendor's equipment; never say that vendor counted those votes.
election_voting_systems
ClaudeFilter the EAC's certified, under-test, Withdrawn or Decertified Systems, terminated and ESTEP e-poll-book rows by list, MORVS-curated vendor label, testing standard or proposed end-of-life notice flag. Defaults to the certified voting-system list. Returns EAC names, standards, dates and source links. A system named in the EAC's April 2026 notice remains on the certified list while the notice is proposed. Federal certification is voluntary; state law decides use. Never call a system insecure or outdated; use decertified only inside the EAC's own list name. Includes the EAC's exact security statement and the November 2024 EAVS survey label.
entity_material_events
ClaudeCross-vertical 'alert me' feed for one entity. Returns only high-signal events (importance ≥ 60): Class I FDA recalls, Schedule 13D filings, large insider sales, restatements, large federal contracts, terminated clinical trials, EPA significant-violator status.
entity_relationships
ClaudeCompany relationship graph (shared-officer edges): other companies whose Form-4 filings name the same person(s) as this entity — board interlock / multi-company executives / activist overlap. Accepts a ticker, CIK, or name.
entity_timeline
ClaudeCross-vertical event timeline for one real-world entity — joins SEC filings/insider/Form144/13D + FDA recalls + FDA drug labels + federal contracts into one ordered feed. Use this when an LLM needs the FULL picture of a company across regulatory verticals. Add min_importance to filter to high-signal events only.
epa_facilities
ClaudeEPA-regulated facilities owned by a company. Returns location, NAICS, and compliance status. Set violations=1 to only see facilities with EPA violations.
estimate_filing_dates
ClaudeEstimates each company's next 10-Q/10-K filing date from its past SEC filing cadence. Dates are estimates, not announced filing dates.
fails_to_deliver_search
ClaudeSearch SEC fails-to-deliver by issuer name or ticker. Returns each security's total fail shares, peak single-day fail balance, and days-with-fails over the rolling 12-month window. CUSIP joins to the SEC securities/13F-holdings layer.
fails_to_deliver_top
ClaudeSEC fails-to-deliver leaderboard — securities ranked by total fail-to-deliver shares over the rolling 12-month window. Fails-to-deliver are a settlement-stress / naked-short red flag tracked by retail investors and market-structure researchers.
fair_market_rent_search
ClaudeSearch HUD Fair Market Rents by area/county/state name. Returns place_id (10-digit HUD fips) to pull an area's FY2026 FMRs — the 40th-percentile gross rent by bedroom size (efficiency/0BR through 4BR). Sets Section 8 voucher payment standards. Aggregate dollar amounts — zero PII.
fcc_grantee_search
ClaudeSearch FCC equipment-authorization grantees — the manufacturer/importer behind every FCC ID — by company name. ~50k grantees across 111 countries (China 12,962, Taiwan 3,717). A supply-chain lens that joins UFLPA forced-labor + trade-screening by name. Optional country filter.
fda_firm_dossier
ClaudeAll FDA activity for one firm — recalls + 510(k) clearances + FAERS adverse-event reports.
fed_enforcement_recent
ClaudeFederal Reserve banking enforcement actions (consent orders, written agreements, CMPs, removals). Optional bank filter.
federal_file_search
ClaudeSearches the MORVS Federal Files Register of federal Privacy Act systems of records by system name, number or ID, with agency, recipient, record-category, status and exemption filters. Recipient tags or groups and record flags are pattern matches on the agency text and count at least those matches; recipient matches exclude standard agency boilerplate. DHS matches do not imply ICE; ICE has its own tag. A routine use is the agency statement that it MAY disclose records for a purpose; it does not show that a disclosure happened. An exemption claim may limit access or correction but does not decide a request. Missing systems are not proof no records are kept. Results default to active systems and include published source context.
federal_file_system
ClaudeReturns one published federal system of records by exact Federal Files Register ID, with the agency's record description, routine-use text, exemption claim, source record and linked Federal Register notices. Standard agency boilerplate routine uses are omitted unless include_standard is true. A routine use is the agency statement that it MAY disclose; it is not evidence a disclosure occurred. A later notice may change a system without reprinting its full text, so check linked notices and effective dates. Recipient tags and record flags are text matches counted as at least; DHS does not imply ICE. Exemption claims may limit access or correction but do not decide an individual request. A missing ID does not establish that no records are kept.
flagged_carriers
ClaudeFMCSA carriers flagged in one or more SMS BASIC safety categories (over the federal intervention threshold): Unsafe Driving, Hours-of-Service Compliance, Driver Fitness, Controlled Substances/Alcohol, Vehicle Maintenance. Returns carrier name, location, fleet size, which BASICs are flagged, and out-of-service inspection counts. ~1,526 carriers currently flagged. Filter by state, a specific BASIC, or minimum alert count.
forced_labor_company_check
ClaudeChecks a company or facility name against the MORVS U.S. Forced Labor Company Register: every organisation named on the DHS UFLPA Entity List or in a CBP withhold release order or finding, the Consolidated Screening List (Commerce, Treasury) rows that match one of those names exactly, and issuer relationships a US-listed company states in its own SEC filing. Matching is exact on the register's name key (case, accent, punctuation and legal-suffix insensitive) or on a source-stated or reviewed alias (DHS "also known as" / "formerly known as", CBP "a/k/a"); there is no fuzzy, family-name or transliteration match, and several matching entities are returned as separate candidates, never merged. Each match carries its direct records, one per source row, with the list's own action, status and date, the source URL, the row locator and the snapshot date; with include_related (default true) it also carries filing-stated issuer relationships, which never extend an action to a parent. UFLPA listing, a CBP order or finding, and an export-control or sanctions row are separate actions under separate laws, and a non-match is not a clearance finding. Organisations only; queries shaped like a personal name are not run.
forced_labor_source_compare
ClaudeCompares two sources of the MORVS U.S. Forced Labor Company Register: the DHS UFLPA Entity List (dhs), CBP withhold release orders and findings (cbp), and the Consolidated Screening List rows that exactly match a DHS or CBP name (csl). Returns the entity ids named in both sources under the register's reviewed name and alias joins, the ids in source_a with no reviewed match in source_b and the reverse, with counts, the register's own published overlap figures, the source rows behind each listed entity (source URL, row locator, snapshot date) and each source's snapshot date, in pages of up to 50 ids per list continued with next_cursor. Issuers linked only through a filing are never counted, so a parent never enters an overlap. The sources apply different laws; an entity with no reviewed match in the other source reads "No reviewed name or alias match in this source snapshot", which is not a clearance finding.
foreign_agents_country
ClaudeLook up a country or location in the Foreign Agents Register by a substring of the country string exactly as DOJ prints it (as entered by registrants and never re-mapped, for example "KOREA, SOUTH"), optionally for one year (2024 or 2025; both when omitted). Returns up to 3 matching countries, each with the money agents reported receiving for its principals for the registrants' six-month periods ending in that year, as DOJ's reports print it, with DOJ's posting date; entries, entries with an amount, "None Reported" entries (never zero) and entries with no Finances line; registrants with money; the withheld part of the country total, as counts and money only; active principal rows and registrants from DOJ's bulk files; the Nature of Services of the published entries; and the largest published registrants and principals with their amounts. Every answer carries DOJ's footnote that country or location is not reflective of U.S. foreign policy and does not imply a recognition of sovereignty. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individuals are withheld: counted only, never named.
foreign_agents_enforcement
ClaudeDOJ's own FARA administration and enforcement figures from its four semi-annual reports to Congress for 2024 and 2025, by half-year, exactly as the register publishes them: books-and-records inspections, civil causes of action, the criminal count with DOJ's own unit and measure, DOJ's enforcement sentences quoted verbatim, and the Statistical Summary (registration statements, terminations, registrants and foreign principals on file, short-form statements, agreements, fees and advisory opinions), with each report's posting date and the statistics note: DOJ's counts differ from the entries its listing prints and are shown as DOJ's, never reconciled. No person is named, and no motive or policy reason is given or implied. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". These are DOJ's counts, a separate measure from FARA receipts, LDA filing counts and PAC contributions, and never added to them.
foreign_agents_principal
ClaudeLook up a foreign principal in the Foreign Agents Register by name substring, as filed or as rewritten to the institution or office where the filed string names a person, optionally narrowed by a substring of the country string as DOJ prints it. Returns up to 5 published principals, each with its type, its country as DOJ prints it, its active rows in DOJ's bulk files, and the published registrants representing it, with the money each reported receiving for or in the interests of that principal for the registrants' six-month periods ending in 2024 and in 2025, as DOJ's reports print it, with DOJ's posting dates. "None Reported" stays its own state, never zero. Entries by withheld registrants count only inside the country totals. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". A registration names who the agent represents, not what any official did. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individual foreign principals are withheld: counted only, never named or matched.
foreign_agents_registrant
ClaudeLook up a registered foreign agent in the Foreign Agents Register by registrant name substring, among the published registrants (reviewed institutions active in DOJ's FARA bulk files or listed in DOJ's 2024 and 2025 reports). Returns up to 5 registrants, each with its type, registration and termination dates and status, its active foreign principals from the bulk files (published ones named, withheld ones as a count), and three federal records side by side: the money it reported receiving for its own six-month periods ending in 2024 and in 2025, as DOJ's reports print it, with DOJ's posting dates, by Nature of Services and by principal (DOJ); its Lobbying Disclosure Act filing count for 2025, or "not searched" (lda.gov); and its firm PAC's status and contributions to federal candidates' committees by cycle, with the FEC definitions (FEC). FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. The columns are never read as a sequence. A firm's PAC gives for the whole firm; nothing in the FEC records ties a contribution to any foreign client. No PAC or LDA match is a search result, not a fact about the registrant. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. Individuals and single-person firms are withheld and never named or matched.
foreign_agents_summary
ClaudeTotals from the Foreign Agents Register, built from the Justice Department's FARA records: the money registered foreign agents reported receiving for the registrants' six-month periods ending in 2024 and in 2025, as DOJ's four semi-annual reports to Congress print it, with each report's printed total, DOJ's posting dates, the basis and the change between the two years; amounts DOJ lists as received prior to registration, kept apart and never added to a period; the share for services DOJ lists as lobbying and the full Nature of Services breakdown by year; active registrants and foreign principals by DOJ's own count for each half-year and by DOJ's daily bulk files; the register's eight headlines with their basis; the side-by-side PAC and LDA coverage counts; and the withheld counts. Registration under FARA is a legal disclosure, not a finding of wrongdoing; in the Justice Department's words, registration "does not indicate approval by the United States Government of the activities of registered agents". Amounts are what agents reported receiving for their own six-month periods, as DOJ prints them, never a calendar year of spending; "None Reported" is its own state, never zero. Country strings are DOJ's, as entered by registrants, and are not a statement of U.S. foreign policy. FARA receipts, LDA filing counts and PAC contributions are separate measures, shown side by side and never added. Individuals (individual registrants, single-person firms, individual foreign principals and short-form registrants) are withheld: counted only, never named.
forfeiture_agency
ClaudeFind DOJ equitable-sharing agencies in one state or territory by agency-name substring or ORI. Returns DOJ database status and sharing labels, matched DOJ payments by federal fiscal year, and the agency's ESAC certifications by its own fiscal year with DOJ and Treasury as separate money fields. DOJ payment years and agency certification years must never be added, subtracted or treated as a gap. A certified balance is not misconduct. DOJ does not publish why an agency has a Non-Compliant or Ineligible status. A payment, review or audit finding does not establish unlawful conduct. Controlled-equipment entries itemise purchases already in ESAC spending categories and must never be added again. A seizure is not a crime or conviction. The newest ESAC year is partial; older thin years are marked in the response. Agency names are DOJ's; officials and other people are never named.
forfeiture_payments
ClaudeSearch DOJ equitable-sharing payments to state, local and tribal agencies by federal fiscal year. Defaults to the latest published DOJ year. Filters by two-letter state, DOJ-printed recipient-name substring and minimum row amount. Returns DOJ cash, sale proceeds, total, original recipient name, match method and matched ORI when one exists. A DOJ-masked recipient stays masked; an unmatched row is still a payment. With no state filter, returns all state subtotals and a separate foreign-sharing table. Foreign sharing is never added to payments to domestic agencies. Agency ESAC certifications use each agency's own fiscal year and include Treasury sharing; never add or subtract them from these DOJ payments. A seizure is not a crime or conviction; a payment follows participation in a forfeiture, not a finding of unlawful conduct.
fund_holdings
ClaudeWhat a mutual fund holds (NPORT-P). Pass a 10-digit CIK or series_id (S000012345).
house_price_index_search
ClaudeSearch the FHFA House Price Index by geography name (metro, state, census division). Returns the place_id to pull a full house-price appreciation history (1975-2026, purchase-only / all-transactions / expanded-data). Aggregate federal statistic — zero PII.
inaugural_donors_donor
ClaudeLook up an organisation donor in the Inaugural Donors Register by name substring, matched against the reviewed public name and every name as filed on the committees' Form 13 reports, optionally for one inaugural year (2025 or 2021; both when omitted, never added across years). Returns up to 5 named organisation donors, each with its reviewed type and, by year, donations accepted, refunded and net (each labelled), every donation row with its date, amount, memo and filing as filed, refunds kept apart (the filing gives no reason for any refund) and the latest filing date. For 2025 donors: the organisation's own LD-203 figure where an item was found and matched, with both amounts and, where they differ, the line that neither record explains the difference; and, for the donors in the register's side-by-side table, three more federal records beside the donation: contract obligations by fiscal year under reviewed parent records (USAspending), the connected PAC's status and its contributions to federal candidates' committees by cycle with its all-committees summary line, two measures never added (FEC), and 2025 lobbying report counts and amounts (LDA). Inaugural donations, contract obligations, connected-PAC contributions and lobbying reports are separate federal records, shown side by side and never added or put in sequence. A missing PAC, contract or lobbying record is a search result, not a fact about the organisation. Committees are named exactly as the FEC lists them, with their ids. Every answer carries the register's question note: the records describe donations reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld organisations (personal trusts, holding vehicles, businesses named for a person, names resting on initials) appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)).
inaugural_donors_summary
ClaudeTotals from the Inaugural Donors Register, built from the Form 13 reports the 2025 presidential inaugural committee (Trump Vance Inaugural Committee, Inc., FEC C00894162) and the 2021 committee (PIC 2021, Inc., C00765040) filed with the FEC: each committee's cumulative printed totals, with donations accepted, refunded and net each labelled and kept apart, the filing ids and filing dates of the latest report and supplement and the coverage dates; organisations named against individual rows, individuals counted only; organisation money by reviewed type; the withheld categories as counts and amounts; the company families that gave in both 2021 and 2025, with their 2021 donations and share, never added across years; the LD-203 cross-check totals (registrants, items, exact, within $1 and differing counts, and each difference with both amounts and the line that neither record explains it), with the in-house registrants for whom no inaugural item was found given as totals only, never by name; the side-by-side coverage counts; and the register's headlines with their basis. Inaugural donations are not campaign contributions. Every answer carries the register's question note: the records describe donations reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Inaugural donations, contract obligations, connected-PAC contributions and lobbying reports are separate federal records, shown side by side and never added or put in sequence. Individual donors are counted, never named; withheld organisations (personal trusts, holding vehicles, businesses named for a person, names resting on initials) appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)).
income_by_zip_search
ClaudeSearch IRS SOI income statistics by ZIP code or state. Returns the 5-digit ZIP to pull its income/tax profile — returns, AGI, average AGI per return, wages, capital gains, EITC, charitable giving by 6 AGI brackets (TY2022). Aggregate federal statistic — zero PII.
influence_chain
ClaudeFollow-the-money chain for one entity: federal contracts + prime->sub subaward flow + registered lobbying + FEC political committees + FEC enforcement (MURs), with dollar tallies. Accepts a company name or a 12-char UEI.
judgment_fund_payee
ClaudeFinds named organisations by payee-name substring in the MORVS Judgment Fund Register. Returns up to 50 matches by published amount, with total match count, summed bi-weekly payment-line amount, line count, agency groups, Treasury citation descriptions, and first and last report periods. Only payees confirmed as organisations appear; individuals are never named. A payment is not an admission of fault, a line is not a settlement, and the separate annual and bi-weekly Treasury products must never be added.
judgment_fund_payments
ClaudeSearches Treasury bi-weekly Judgment Fund payment lines in one fiscal year of the MORVS Judgment Fund Register. Defaults to the latest published fiscal year, which may be incomplete. Filters by named organisation payee, Treasury responsible agency or MORVS agency group, and minimum known grouped amount. Returns separate source lines grouped by named payee, agency and report period; unnamed lines remain separate. Includes payee status, Treasury report links and source row numbers. Individuals are never named. A payment is not an admission of fault; a line is not a settlement, and split or identical lines remain separate payments. The bi-weekly and annual Treasury products are separate and must never be added.
kev_by_vendor
ClaudeCISA Known Exploited Vulnerabilities (KEV) for one tech vendor. Returns CVE list + ransomware count + top affected products. The federal government's canonical list of CVEs being actively exploited.
kev_recent
ClaudeRecent CISA KEV entries. Filter by vendor or ransomware-known flag.
lobbying_by_client
ClaudeAggregate Senate Lobbying Disclosure dossier for one company: total spend, firms used, top issue codes, recent filings.
lobbying_by_issue
ClaudeWho lobbies the federal government on a given issue. Takes an LDA issue code (e.g. HCR health, DEF defense, TAX, ENG energy, TRA transport) and returns top clients by reported spend. List codes at /api/v1/lobbying/issues.
lobbying_client
ClaudeReturns client timelines from the MORVS Influence Register using a published key or a case-insensitive name substring. The register includes only reports with an AI, surveillance or privacy match in a lobbying-activity section; it is not a client’s complete lobbying history. Amounts are whole-report figures across all issues, never spending on a topic. In-house expenses and outside-firm income are separate and must not be added; subcontract income is excluded. Null is not zero. Filings do not establish a client’s position on a bill or a policy outcome. Returns the retrieval date and the LDA source disclaimer. A missing client is not proof of no lobbying.
mdl_lookup
ClaudeFind a federal multidistrict litigation (MDL) in the MORVS Mass Lawsuit Register by MDL number, caption word, or company word appearing in the Panel caption. Returns the Panel caption, category, district, transfer date, pending and total actions, the last 12 monthly pending counts, and terminated entries when available. Actions are lawsuits, not people. Centralisation is not a finding on the merits. A falling pending count is not a settlement or payout. A missing match only means no matching MDL was found in the published register.
migrant_programs_program
ClaudeOne program of the Migrant Program Money Register by key: its label, agency and award-ID rule (the FAIN prefix pattern and the listings it is booked under, or the reviewed award list for contracts) with the register's evidence; the listing note where the program's money moved between assistance listings (the unaccompanied-children program booked under 93.566 from December 2024 and re-booked older obligations in September 2026, so a series by listing alone moves with that booking), and money by listing or by kind where the register gives it; by fiscal year, net obligations, positive obligations, de-obligations, transactions, recipients and new awards, each labelled with its fiscal year and the as-of date; the 10 largest recipients each fiscal year as the register ranks them, with their class; new awards in FY2025 and FY2026 by funding opportunity; the equal twelve-month windows; class shares by year; and, for State Reception and Placement, the awards placed outside the United States that are excluded. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts.
migrant_programs_recipient
ClaudeLook up a recipient in the Migrant Program Money Register by name substring as the award records name it, by the register's family name, or by its exact 12-character UEI, optionally limited to one program. Returns up to 5 published recipients, each with its UEI, states, reviewed class and the basis for it, and the IRS Exempt Organizations Business Master File record where one matches (EIN, the name and state the IRS lists, and the IRS's own codes); money by program and fiscal year; its 10 largest awards with obligations by fiscal year and award-level outlays as recorded, each labelled; its termination records in the agency's exact words with their dates; the subaward rows its prime awards report, as counts only (the register gives no subawardee total); and the register's side-by-side columns for its family: program money, LDA lobbying reports, connected PACs, ICE and CBP contracts and super PAC gifts, with the caption. An empty column is a search result, not a fact about the organisation. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing.
migrant_programs_summary
ClaudeTotals from the Migrant Program Money Register, built from USAspending award records, HHS's TAGGS terminated-grants list and the IRS Exempt Organizations Business Master File: federal obligations, FY2023 to FY2026, to the organisations and governments that shelter, care for, resettle, transport and give legal help to migrants and refugees, in eight programs keyed from the award ID (ORR unaccompanied-children grants and ORR contracts, ORR refugee formula, Matching Grant and discretionary grants, State Department Reception and Placement, FEMA's Shelter and Services Program and DOJ EOIR legal access). Returns the all-programs total with its basis; money by program and fiscal year, with net obligations, positive obligations and de-obligations labelled; the equal twelve-month windows either side of January 20, 2025; recipient-class shares by program and year, with state, local and tribal governments apart from organisations and organisations split by IRS match; FEMA's Shelter and Services de-obligations of February 10 to 12, 2025 (the record gives no reason), with the government share; termination records counted by agency and record type; withheld recipients and records as counts and amounts; the exclusions, the awards placed outside the United States listed with their amounts; and the register's headlines with their basis. Awards whose record places the work outside the United States are excluded from every figure, as are FEMA's Emergency Food and Shelter Program and ICE detention money. Every amount is an obligation from the award records (the government's commitment to pay), net of de-obligations, by action date; net obligations, positive obligations and de-obligations are separate stages, each labelled with its fiscal year and the as-of date, award-level outlays (only where the record carries them) are the record's cumulative figure to the as-of date, and they are never added across stages. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Lobbying reports (LDA), connected PACs (FEC), ICE and CBP contracts (the Pay-to-Play Register, where ICE detention sits) and super PAC gifts (the Super PAC Donors Register) are side columns: shown side by side, never added to program money and never put in date sequence with an award or a termination. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts.
migrant_programs_terminations
ClaudeList the termination records in the Migrant Program Money Register, optionally filtered by agency (a name substring as the record or its program names it, or a short form: FEMA, DHS, USCIS, ICE, CBP, HHS, ACF, ORR, State, PRM, DOJ or EOIR, each resolved to the agency names the data gives; a short form the data does not name returns no rows and the answer lists the agency names in the data), program key and recipient name substring; at most 25 rows, with the full count and the count by record type. Three record types, each verbatim with its date and amount as the record gives it: rows of HHS's TAGGS terminated-grants list, in HHS's own words and columns (total obligated, expended, payment and unliquidated, labelled as HHS labels them) with the list's date; termination and recipient-request texts from the USAspending award records, with the agency's exact words, the action type as filed, the action's obligation and any executive-order tag exactly as the record shows it; and FEMA's Shelter and Services de-obligations of February 10 to 12, 2025, by recipient, each labelled the record gives no reason. A $0 termination action is not the money: any de-obligation is a separate, later action. Terminations appear only in the agency's exact words with their dates; where the record gives no words the answer says the record gives no reason, and an executive-order tag appears only as the record shows it. Governments are kept apart from organisations: every amount names its recipient class from the register's reviewed table, never a single NGO total. A recipient is called an IRS-listed exempt organisation only where the register's IRS EO BMF match stands (by the reviewed EIN, or by exact name and state where the register marks the match by rule), shown with that record; other organisations read organisation, IRS status not matched. The tools never call a recipient nonprofit or faith-based, and no recipient is classified by religion. Every answer carries the register's question note: these records show federal awards and the agencies' own termination records; they do not show how any recipient used the money or why an award was made or ended beyond what the record says. Every answer also carries the side-by-side caption, the as-of and retrieval dates, each file's date, the source page on morvs.ai and the licence. Institutions only: no beneficiary, staff member, executive or sole proprietor is named; withheld recipients and records appear only as counts and amounts. A name that matches only a withheld string answers exactly like a name that matches nothing.
migration_flows_search
ClaudeSearch IRS county-to-county migration flows by county/state name. Returns place_id ({state_fips}-{county_fips}) to pull a county's net migration (returns/people/AGI in vs out, FY2022-2023) and its top origin/destination counties. Aggregate IRS statistic — zero PII.
national_security_surveillance
ClaudeReturns the published agency counts in the MORVS Surveillance Orders Register national-security source ledger. Optionally filter by exact series ID and period. Every fact retains the agency label, source, page, quote, method and any qualifier or disagreement. Section 702 targets, FISA orders, national-security letters and border device searches use different units, laws and periods; never add them or treat them as one trend. Section 702 target counts estimate non-US persons abroad; FBI query counting changed. These are aggregate published counts and cannot identify surveillance of a person.
news
ClaudeTime-ordered cross-topic news feed: insider trades, Form 144 notices, cluster buys/sells, Schedule 13D filings, 8-Ks, FDA recalls, weather alerts, presidential documents.
nhtsa_automation_investigations
ClaudeSearch MORVS-reviewed NHTSA automated-driving, Level 2 driver-assistance and SGO-compliance investigations. Returns NHTSA’s structured action fields, cited SGO report IDs and published lookup links; scope and reason are MORVS-reviewed classifications, not NHTSA findings. Optionally includes linked recall campaigns with potentially affected units as filed by the manufacturer. An opened investigation or a defect petition is not a finding. A recall’s potentially affected units are not a fleet census. No NHTSA summary text or personal names are returned. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
nhtsa_by_manufacturer
ClaudeNHTSA vehicle recall dossier for one auto OEM. Returns total recalls, top affected components, park-it count, recent recalls. Public auto OEMs only.
nhtsa_recent
ClaudeRecent NHTSA vehicle recalls. Filter by manufacturer, make, model year. park_it=1 returns only the most urgent recalls (do-not-drive).
nih_by_org
ClaudeNIH RePORTER dossier for one organization — total federal R&D funding received, top activity codes, recent grants. Best for academic medical centers + biotechs.
nsf_by_org
ClaudeNSF Awards dossier for one organization — total funding, top programs, recent awards. Pairs with academic medical centers + biotech/tech startups.
patents_by_assignee
ClaudeAggregate USPTO patent dossier for one company: total grants, top CPC classes, recent grants, inventor counts.
patents_recent
ClaudeRecent USPTO patent grants from PatentsView. Filter by assignee company or CPC class. Use for R&D intel and IP portfolio questions.
pay_to_play_awards
ClaudeThe largest ICE and CBP awards in the Pay-to-Play Register: awards funded by the component with a transaction in FY2025 or FY2026, ranked by Award Amount. Filter by component (ICE or CBP), fiscal year (2025 or 2026), recipient name substring, or the law File C records for the award (PL119-21, PL119-98 or account_level_only); limit 1-50 (default 20). Each award gives the recipient, its parent, the Award Amount and outlays (both lifetime figures), the award description verbatim from the award record, and the File C law link with its class (labelled or probable) or "account level only"; total and truncated give the full count. Award Amount is the award's lifetime amount, not an FY2025-26 obligation, and is never added to the fiscal-year contract totals. A File C law link records the account the award was charged to, not a judgment about the award.
pay_to_play_committees
ClaudeList the federal candidate committees that received contributions from the connected PACs of the 100 largest ICE and CBP contractors in the Pay-to-Play Register, for one FEC cycle (2024 for 2023-24, or 2026 for 2025-26, default 2026). Filter by committee name substring, state (two letters), party (REP, DEM or other, as the FEC lists it) or company name substring; limit 1-50 (default 20). Each committee name is as the FEC lists it, with party, office, state and district from the FEC, the net amount from the mapped PACs and the split by company; total and truncated give the full count. A listing records a contribution reported to the FEC; it is not a finding about the candidate, the contribution's purpose or any vote. No individual donor is named and no candidate profile is built. PAC money is company-wide, not tied to ICE or CBP work, and comes from voluntary contributions, not company treasury or contract money; the FEC's own words are quoted.
pay_to_play_contractor
ClaudeLook up an ICE or CBP contractor in the Pay-to-Play Register by company name substring, matched against the company, the parent as USAspending lists it, the reviewed alias note and its recipient names, among the 100 largest parents by FY2025-26 contract obligations. Returns up to 5 parents, each with three federal records side by side: ICE and CBP contract obligations by component and fiscal year (funding view, with the awarding view beside it), share, File C obligations on the border wall system lines and recipients (USAspending); connected-PAC status, committees as the FEC lists them and, by cycle, the PAC's contributions to federal candidate committees and its summary line for contributions to all committees, two measures that are never added (FEC); and lobbying reports by year, with in-house expenses and outside firms' income kept apart and never added, and IMM-coded counts (LDA). The three columns are never added together and are never a sequence. The register's note: "The register's name labels a question people ask; it is not a finding. These records do not show that any contract was exchanged for money, or that any contribution or lobbying influenced any award." PAC totals are company-wide, not tied to ICE or CBP work, and come from voluntary contributions, not company treasury or contract money; the FEC's own words are quoted. The absence of a PAC or LDA filing is a search result, not a fact about the company. A company outside the 100 largest parents is a scope statement, not an absence.
pay_to_play_totals
ClaudeICE and CBP contract obligation totals from the Pay-to-Play Register, by component (ICE, CBP or both, default both) and fiscal year (2025, 2026, or both when fy is omitted): the funding view (contracts funded from the component's own appropriations, the headline measure) with the awarding view and the bridge between the two beside it, ICE and CBP combined in the funding view, and the share held by the 5, 10 and 25 largest parents. Also the P.L. 119-21 border wall system (Sec. 90001, CBP) by stage: appropriated, obligated and outlaid, each with its own as-of date; the two laws' lines by program activity, labelled where the program activity's name cites the law and shown as probable where only the account's availability window matches, never counted as the law's money; and the register's eight headlines with their basis and as-of dates. Appropriated, obligated and outlaid are separate stages and are never added. Contract obligations are orders placed and contracts signed, not outlays (money paid out). Every answer carries the register's question note and caption: the records are shown side by side and do not show that one caused another.
pfas_rule_impact
ClaudeCompares two rule scenarios in the MORVS PFAS Drinking-Water Register: 2024_as_published (EPA's six 2024 PFAS drinking-water limits, final and in force) and if_proposals_final (the PFOA and PFOS limits only, as they would stand if EPA's May 2026 proposals to rescind the PFHxS, PFNA, HFPO-DA and Hazard Index limits became final; they are proposals, not final rules). Returns each rule record with its status label, the number of systems and the ECHO population served with at least one UCMR 5 sampling-location average at or above a scenario limit's comparison cutoff (EPA Table 4 method), the PWSIDs whose only such averages are against limits proposed for rescission, the largest matching systems by population up to limit, the register's published figures for comparison and the method version, optionally filtered by state file, owner type and minimum population. These are technical comparisons of monitoring averages, not compliance results (not_compliance: true).
pfas_system_history
ClaudeReturns one public water system's full record in the MORVS PFAS Drinking-Water Register by PWSID: each sampling location (a hashed key) with its sample events, first and last dates and complete-set averages in ppt, per-chemical sample and detection counts with the highest single result, the Hazard Index, the six derived flags, the six 2024 rule records with each one's status history and whether this system's averages meet its comparison cutoff, the source file hashes and the register's caveats. Raw samples are not published. UCMR 5 results are entry-point monitoring, not compliance determinations or household tap results, and EPA's May 2026 rescission and deadline documents are proposals. A system whose name the register withholds is shown by PWSID and city.
pfas_system_lookup
ClaudeLooks up public water systems in the MORVS PFAS Drinking-Water Register, built from EPA's final UCMR 5 monitoring file, the ECHO SDWA system file and the Federal Register. Give exactly one of pwsid (one system), zip (every system that reported serving that ZIP code; a service-area hint, not a confirmed supplier) or state (one state file, or TRIBAL), optionally with name to filter a state's systems by system name or city. Each match carries the PWSID, display name, state, city, ECHO population served, owner type and its label, activity status, the six derived flags, the 2024 limits whose comparison cutoff at least one sampling-location average meets, and for ZIP queries the match scope; the answer carries source dates, the status of each limit named and the register's caveats. UCMR 5 results are entry-point monitoring samples compared with the 2024 limits by EPA's Table 4 method, not compliance determinations or household tap results, and EPA's May 2026 rescission and deadline documents are proposals. A system whose name the register withholds is shown by PWSID and city.
pharmacy_340b_clients
ClaudeReverse 340B lookup: given a contract-pharmacy ID, list every 340B covered entity that dispenses its discounted drugs through that pharmacy — reveals how concentrated the contract-pharmacy network is around big chains. Pharmacy IDs come from a covered entity's detail page.
plate_reader_decisions
ClaudeFinds official city and county automated license plate reader matters in the MORVS License Plate Reader Register. Returns selected published decision fields with complete recorded action timelines, source links, vendors named in titles, and dollar figures exactly as printed. Filter by state, jurisdiction, vendor, recorded motion outcome, subject, decisive-action year or title substring query. Outcome means only the motion as recorded. An approval is an authorization, not proof of a signed contract or installed camera. A failed motion does not mean a contract ended. Amounts are printed figures, often ceilings, never money spent. No row means no qualifying action was found in the checked sources and period; it is not proof that no ALPR use exists.
plate_reader_state_law
ClaudeReturns the MORVS License Plate Reader Register's full state-law research record for one US state or DC, including official law citations and URLs, checked date, verification status, retention and sharing provisions, exceptions and search limitations. A "none_found" or unverified status describes the sources checked; it does not prove that no applicable law exists. Verify the current official text before relying on a provision.
police_grant_awards
ClaudeSearch the Police Grant Register's DOJ COPS award lists by fiscal year (FY2020–FY2026, default latest), program, state and recipient name. Amounts are DOJ's announced awards, not outlays; the lists do not record scoring, so an award must never be attributed to a recipient's immigration posture. DOJ's Estimated Award Amount columns are labelled estimated. The FY2024 technology rows are held because that list does not reconcile; DOJ's printed total is identified as its figure only. JAG formula allocations, COPS awards and forfeiture payments are never added together. Reviewed DOJ-list and ICE 287(g) matches are labelled as such; forfeiture payments appear only side by side.
police_grant_recipient
ClaudeFind institutional recipients by name substring in the Police Grant Register's DOJ COPS award lists for FY2020–FY2026. Returns each recipient's award rows by program and fiscal year, reviewed DOJ-list and ICE 287(g) match labels, and any DOJ equitable-sharing payments side by side. COPS, JAG formula allocations and forfeiture payments are never added. 'No COPS award appears in lists for FY2020–FY2026' never means an agency gets no federal money; absence covers only the published COPS lists. Award lists do not record application scoring or money paid out.
prediction_market_exchange
ClaudeFind a CFTC designated contract market (DCM) or pending application by exchange code or name substring in the MORVS Prediction Market Register. Returns CFTC-listed status/date and fixed-phrase remark flags, filing counts by listed year/category/status, stay and withdrawal counts, approval-pending and perpetual-futures counts, code-map evidence and open federal comment periods. Pending means only pending; a stay is a CFTC remark, not a finding. Filing counts measure filing practice, not markets, trading, users or odds. Dates are as listed by the CFTC and may be re-dated.
prediction_market_filings
ClaudeSearch event-type CFTC product filings in the MORVS Prediction Market Register by exchange code/name, listed year, filing status, MORVS topic, CFTC stay remark or public title substring. The default year is the latest published listed-year shard. A Certified filing was self-certified by the exchange under CFTC rule 40.2; it was not approved by the CFTC. Approval Pending and Approved are a separate product-approval route. Filings are not markets, volume, users or odds. Dates are as listed by the CFTC and can be re-dated after a status change. Topics are MORVS classification. Titles that may name individuals are withheld and are never searched; use the CFTC ID link for the source record. A stay is not a finding of a violation.
resolve
ClaudeEntity locator: resolve a company name/ticker/CIK to its canonical SEC identity + aliases, and an available_in map of which federal datasets contain it (with links). Use first to find an entity and discover where it has data.
sanctions_check
ClaudeSearch MORVS's partial local OFAC SDN index by case-insensitive substring, including indexed aliases. Returns hit=true for a local match. A non-match is not sanctions clearance; check /coverage and OFAC's current official lists.
sanctuary_doj_list
ClaudeSearch DOJ’s current list page or its August 2025 announcement by name; defaults to the current page. “Sanctuary jurisdictions” is DOJ’s label, not a finding by MORVS. The two DOJ pages differ; removed entries carry DOJ’s linked release and stated reason. A county-level 287(g) agreement and a state-level designation can both be true; no conclusion is drawn from the two records. A non-match says only that a name was not found on the selected DOJ page; DOJ says its list is not exhaustive.
sanctuary_state_reports
ClaudeSearch published California Values Act transfer counts, Illinois Way Forward Act detainer, warrant, transfer and request reports, or Oregon Sanctuary Promise Act requests and communications. These are agency self-reports. State measures differ and are never added or ranked across states. State is required; agency name, year and result limit are optional. Oregon’s year is the end year of its June–May reporting period. Illinois agency rows are available for 2024; 2023 has statewide figures and a separate non-filer list, not agency appendix rows. A zero is a reported zero, a null is absent from an appendix. “No transfers reported” and “did not file” do not mean “no cooperation”. Never say an agency “violated” a law or acted “illegal” based on these records.
screen_all
ClaudeExperimental lookup of one name across MORVS's local public-record indexes; availability is not guaranteed. Returns per-dataset match counts and an indicative 0-100 score (risk_score) summarizing how many indexed records matched. The score is not a risk rating or compliance determination, and a low score or non-match is not clearance. Coverage and freshness vary by dataset.
screen_batch
ClaudeScreen up to 50 entities in one call. Returns a per-entity risk_score + risk_level + per-vertical summary for each, plus the list sorted highest-risk-first. Use to vet a whole vendor / counterparty / portfolio list at once.
search
ClaudeFull-text search across all parsed SEC/FDA/register/weather section bodies (BM25 ranking). Also resolves company names for ticker lookup.
search_340b
ClaudeSearch HRSA 340B Drug Pricing Program covered entities (disproportionate-share hospitals, FQHCs, clinics buying federally-discounted outpatient drugs) by name. Returns 340B ID, entity type, state, participating status, and contract-pharmacy count. ~93k entities.
self_driving_company
ClaudeFind reporting companies in the MORVS Self-Driving Crash Register by published name or slug. Returns separate ADS and ADAS company profiles, report and incident counts, severity, location, confidentiality index, investigations, recalls and California DMV permits. share_of_kind is a share of reports, not a crash rate. A DMV permit holder may be a different legal entity; the register notes that difference. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
self_driving_reports
ClaudeSearch NHTSA Standing General Order crash-report rows in the MORVS Self-Driving Crash Register. Requires ADS or Level 2 ADAS and can filter by company, state, alleged injury severity, crash month and the year NHTSA first received the report. Returns latest report versions with NHTSA CSV citations, not narrative text. No report has its own NHTSA web page. OTHER filings are never returned as crashes. Reports are each company’s own filings with NHTSA, not a finding, and say nothing about responsibility. The highest injury severity alleged says nothing about responsibility. ADS and Level 2 ADAS have different reporting requirements, are kept separate and are never added together; OTHER is never a crash count. Reports and crashes are different counts. Counts reflect fleet size and reporting rules; they are not rates, are never normalised by miles or fleet size, and are never a safety ranking. The third amended order changed reporting from June 16, 2025, so counts across that break are not directly comparable. A narrative marked as confidential business information is a claim, not an NHTSA decision. No people or narrative text are returned. NHTSA’s caveats are included verbatim.
short_interest_search
ClaudeLook up a ticker's FINRA short interest: shares sold short, days-to-cover, change vs prior settlement. The real short metric (distinct from short volume which includes market-maker hedging). Pairs with fails-to-deliver + short volume. FINRA source terms: https://www.finra.org/terms-of-use.
short_interest_squeeze
ClaudeFINRA equity short interest: securities with the highest days-to-cover (the classic short-squeeze indicator) + the largest absolute short positions, from the latest bi-monthly settlement. TRUE short interest (not short volume). Use to find squeeze candidates and the most-shorted stocks. FINRA source terms: https://www.finra.org/terms-of-use.
short_squeeze_candidates
ClaudeTop short-squeeze candidates by a MORVS (AI Analytics LLC) 0-100 pressure score combining FINRA days-to-cover + short-sale volume % + SEC fails-to-deliver per symbol. The methodology is separate from and does not relicense FINRA inputs. Transparent heuristic, not investment advice. FINRA source terms: https://www.finra.org/terms-of-use. Optional min_score filter.
short_volume_most_shorted
ClaudeFINRA short-sale volume: the most-shorted liquid stocks by average short % of volume (rolling ~60 trading days), plus dataset summary. Pairs with fails-to-deliver for the short-selling picture. Caveat: FINRA short volume includes market-maker hedging, not short interest. FINRA source terms: https://www.finra.org/terms-of-use.
short_volume_search
ClaudeLook up a ticker's FINRA short-sale volume: average / latest / peak short % of consolidated volume over the rolling ~60-trading-day window. FINRA source terms: https://www.finra.org/terms-of-use.
super_pac_donors_donor
ClaudeLook up an organisation that gave to super PACs or hybrid PACs in the Super PAC Donors Register by name substring, matched against the reviewed name, every name as filed and the reviewed company family, optionally for one FEC cycle (2026 for 2025-26, or 2024 for 2023-24; both when omitted, never added across cycles). Returns up to 5 named organisations, each with its reviewed class and status, its company family and sector, and the parent or group record from an SEC filing where the register cites one; and by cycle, its contributions by recipient super PAC as the FEC lists it, with the committee id, refunds kept apart and never netted, dates as filed and the FEC bulk file date. Beside them, never added or put in sequence: for the company families in the register's side-by-side table, the connected PAC's contributions to federal candidates' committees with its all-committees summary line (FEC), contract obligations by fiscal year (USAspending) and LDA filing counts; the organisation's net donation to the 2025 inaugural committee where it is on both records (Form 13); and ICE and CBP contract obligations for Pay-to-Play Register contractors. An empty column is a search result, not a fact about the organisation. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. A name that matches only a withheld or unreviewed string answers exactly like a name that matches nothing. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
super_pac_donors_recipient
ClaudeLook up a super PAC or hybrid PAC in the Super PAC Donors Register by committee name substring, as the FEC committee master lists it, or by its committee id (C followed by 8 digits). Returns up to 5 committees, each for the 2025-26 and 2023-24 cycles with its name, id and committee type as the FEC lists them, and three measures, each labelled and never combined into one figure: organisation contributions by reviewed class (issue and social-welfare groups split by IRS match; withheld and unreviewed money as amounts only), contributions from committees, and itemized contributions from individuals, counted only; conduit pass-through rows (ActBlue, WinRed) are shown apart under their own label and are excluded from every total. Each cycle lists the largest named organisation donors to that committee and the committee's own FEC summary line. Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. Organisation, committee and individual money into super PACs are three measures, each labelled and never combined into one figure. Organisation money is not corporate money: every total states its class. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
super_pac_donors_summary
ClaudeTotals from the Super PAC Donors Register, built from the FEC's bulk files: organisation contributions to super PACs and hybrid PACs for the 2025-26 and 2023-24 cycles, with rows, names as filed, named donors, the reviewed share and each FEC bulk file's date, and the change over the same date window in both cycles; organisation money by reviewed class, with issue and social-welfare groups split between those an exact IRS Exempt Organizations record matches and those whose status was not matched, beside companies, unions, trade associations, committees, tribal governments and other organisations; withheld and unreviewed strings as counts and amounts only; itemized contributions from individuals reported directly to these committees, counted only, beside organisation money on that exact basis, with conduit pass-through rows (ActBlue, WinRed) shown apart and excluded from every total; sectors and the largest recipient committees as the register lists them; the lag line; and the register's headlines with their basis. Organisation, committee and individual money into super PACs are three measures, each labelled and never combined into one figure. Organisation money is not corporate money: every total states its class. Super PACs and hybrid PACs are named exactly as the FEC committee master lists them, with their committee ids and no party or candidate label; a contribution to a super PAC is not a contribution to a candidate. The word nonprofit appears only on a record an exact IRS Exempt Organizations match carries, shown with that record; other issue and social-welfare groups read organisation (status not matched). Every answer carries the register's question note: these records show contributions reported to the FEC; they do not show what any donor received or intended, and they do not show cause and effect. Individual donors are counted, never named; withheld strings (trusts, holding vehicles, businesses named for a person, names resting on initials, unmatched committees) and unreviewed strings appear only as counts and amounts. Every answer quotes the FEC's restriction on the sale or use of individual contributor information (52 U.S.C. 30111(a)(4)) and the FEC bulk file date; the bulk file lags the committees' latest reports.
trade_remedy_search
ClaudeSearch USITC trade-remedy investigations — antidumping/countervailing-duty (AD/CVD) + Section 337 unfair-import (patent/IP) cases — by investigation number, product, or title. ~4,386 cases; the trade-enforcement vertical (companies/countries fighting over dumped, subsidized, or IP-infringing imports).
ukraine_money_recipient
ClaudeFind institutions and companies in the Ukraine Money Register by name substring: named award-record recipients of Ukraine-division accounts (obligated by fund code, obligated and outlaid by class, place of performance) and DoD contracts whose own records mention Ukraine, grouped by parent as USAspending lists it. With include_loop (default true), loop companies show three records side by side, never added and with no causation: DoD Ukraine-mention contract obligations; connected-PAC contributions to federal candidate committees by cycle, with the top recipient committees named exactly as the FEC lists them, with party and office; and lobbying filings citing Ukraine bills, each with its all-issue amount. Corporate PAC money is voluntary contributions from people the PAC may solicit, not company treasury or contract money; the FEC's own explanations are quoted. No individual donors, and no withheld or individual award recipients. The published recipient list holds the largest named institutions only, so a non-match is not evidence that no award was made.
ukraine_money_totals
ClaudeTotals from the Ukraine Money Register of federal Ukraine-response money. view=stages (default): the inspectors general's made available, obligated and disbursed (outlaid) totals as printed, each with its own as-of date; acts: the five supplemental acts' made-available totals re-added clause by clause from the act text, beside SIG-OAR's table; codes: USAspending's emergency fund codes 3, 6, AAA, AAB and AAJ, where the mixed codes are split against the act text into Ukraine-only, probable and shared accounts, shown apart; places: place of performance in the award records; oversight: inspector-general and GAO findings quoted exactly with report numbers, dates and pages. Stages (made available, obligated, outlaid, announced) are never added to one another. Pentagon contracts whose own records mention Ukraine are not tied to the supplemental appropriations, and USAspending links only $0.47 billion of DoD's $121.30 billion obligated under the five codes to named awards. Words alleging loss or wrongdoing, such as fraud or diversion, appear only inside exact quotes.
upcoming_earnings
Claudeupcoming_sales_pressure
ClaudeLists companies by the size of insider sales planned in recent SEC Form 144 notices (filed intentions to sell, not completed trades).
watchlist
ClaudeConsolidated activity rollup for a list of tickers — insider 30d net flow, Form 144 pressure, recent material events, cluster-sell flag. Up to 25 tickers per call.
wiretap_orders
ClaudeFinds individual court-authorized wiretap orders as published in one annual US Courts Wiretap Report. Defaults to the latest report year; filters use the report year, court, state, jurisdiction, offense group and intercept type. A row was reported to the courts in that report year and may have ended earlier. Results and costs are per order as reported. The annual headline count undercounts that year because some orders are reported in later years. Related orders can share people, intercepts and arrests, so do not sum those fields across rows. Zero or blank means none reported, not that nothing happened. Flags identify missing prosecutor reports and implausible figures. A missing row is not proof that no surveillance occurred; the published order shards start in 2016.
