get_agency
ChatGPTFetch one SAM.gov Federal Hierarchy organization by fhOrgId, as most recently ingested. Returns its name, type, status, parent, agency code and CGAC codes. Use when you have an fhOrgId or parentFhOrgId from search_agencies. Prefer search_agencies when you only have a name or code. Example: {"orgId":"300000352"}
get_assistance_listing
ChatGPTFetch one SAM.gov Assistance Listing (CFDA program) by assistanceListingId, as most recently ingested. Returns its title, objective, description, organization codes and Federal Hierarchy match. Use when you know the listing number. Prefer search_assistance_listings when you only have a topic or agency. Example: {"listingId":"10.227"}
get_award
ChatGPTFetch one contract award transaction by recordId. Returns structured columns (PIID, agency, vendor, NAICS and PSC with descriptions, amounts, dates), no raw upstream payload. Use when you need the full record for a row from search_awards. recordId is the opaque base64url value search_awards returns, not a UUID. Prefer get_award_hierarchy when you need the parent IDV or child actions. Vendor fields are null on rows dated before Apr 4, 2022 (D&B-sourced) or with unknown action dates; vendorFieldsWithheldReason identifies withholding, otherwise null. Example: {"recordId":"eyJ2IjoxLCJrIjoiYSIsImZ5IjoyMDI3LCJtIjoiMjAyNi0xMCIsInQiOiJDT05UX0FXRF8zNkMyNTYyN0YwMDQxXzM2MDBfNDdRU0hBMjBEMDAxMV80NzMyIn0"}
get_award_hierarchy
ChatGPTFetch the parent IDVs, child award actions and same-award transactions for one contract award or IDV record. Returns those records plus child totals: summed action obligations, distinct award keys and cumulative contract value counted once per award; limit (default 10, max 25) caps listed items, not totals. Use when asking how an award fits under an IDV or which orders sit under a vehicle. recordId is the opaque base64url value from search_awards, not a UUID. Prefer search_awards when you do not have a recordId yet. Vendor fields are null on rows dated before Apr 4, 2022 (D&B-sourced) or with unknown action dates; vendorFieldsWithheldReason identifies withholding, otherwise null. Example: {"recordId":"eyJ2IjoxLCJrIjoiYSIsImZ5IjoyMDI3LCJtIjoiMjAyNi0xMCIsInQiOiJDT05UX0FXRF8zNkMyNTYyN0YwMDQxXzM2MDBfNDdRU0hBMjBEMDAxMV80NzMyIn0","limit":10}
get_entity
ChatGPTFetch one SAM.gov public registration by UEI. Returns registration fields, addresses, size representations, raw SAM business-type codes and socioeconomicCertificationsJson (SBA-issued certifications only); code labels add primaryNaicsLabel. Use when inspecting raw registration fields. Prefer get_entity_profile when you need plain business-type labels, self-represented programs, connected awards or risk records. Example: {"uei":"P2SNKWSMC831"}
get_entity_profile
ChatGPTFetch one SAM.gov entity registration with NAICS-specific size representations, business types with plain labels (registration.businessTypes), self-represented programs such as SDVOSB/VOSB/WOSB/EDWOSB (registration.selfRepresentedPrograms), dated SBA-issued certification status (registration.sbaCertifications: 8(a), HUBZone, SDB, and SBA-certified WOSB/EDWOSB codes A9/A0; an empty list does not mean no socioeconomic status), recent contract awards, agency/NAICS/PSC rollups (default 5), exclusions and integrity records. Returns one profile object. Use when researching one vendor's eligibility, risk or awards by UEI. Prefer get_entity when you only need raw registration fields, or search_entities when you do not have a UEI. Section limits max 25. Connected awards and rollups exclude rows dated before Apr 4, 2022 and unknown action dates. Example: {"uei":"P2SNKWSMC831","awardsLimit":5}
get_exclusion
ChatGPTFetch one SAM.gov exclusion by recordId. Returns address, classification, exclusion type/program, excluding agency, active/termination dates, comments and join keys (UEI, CAGE, NPI). Use when you need the full record for a row from search_exclusions or list_entity_exclusions. recordId is opaque base64url, not a UUID. Prefer search_exclusions when you only have a name or identifier. Example: {"recordId":"eyJ2IjoyLCJrIjoieCIsInIiOiI2ZDQzYTc3Yy0zYjY3LTQxYmQtYTAzZS03NmE5ODc2Y2ExZjkifQ"}
get_integrity_record
ChatGPTFetch one SAM.gov Responsibility/Qualification or FAPIIS integrity record by recordId. Returns the full row: firm identifiers, record family, dates, agency, title, description, disposition and proceeding/termination details; code labels add pscLabel and primaryNaicsLabel. Use when you need the full record for a row from search_integrity_records. recordId is opaque base64url, not a UUID. Prefer search_integrity_records when you only have a UEI or name. Example: {"recordId":"eyJ2IjoyLCJrIjoiaXIiLCJyIjoidGVybWluYXRpb258NGZlZThkNDc5MjJlYzE0MWNlYTkwMWFhNmFiNTcwMjN8MSJ9"}
get_labor_rate
ChatGPTFetch one CALC+ labor ceiling row by recordId. Returns {data} with every reported field, three exact contract-year ceilings, source load/capture provenance and nullable eLibrary contract/UEI; missing IDs return not-found. Use when you have a recordId from search_labor_rates and need the full source facts. Prices are not invoices or guaranteed order pricing. Prefer search_labor_rates when you do not have a recordId. Example: {"recordId":"gsa-calc:000014304a09d5d62765df715ea8456a8e1af639cb314ca5288a165b3d147142"}
get_mentor_protege_agreement
ChatGPTFetch one published SBA mentor-protégé relationship by recordId. Returns {data} with approval date, both reported firms and identifiers, qualifications, source provenance, nullable exact company references and identifier statuses; missing IDs return not-found. Code labels add naicsLabel. Use when you have a recordId from search_mentor_protege_agreements. Fallback IDs can change after source identity repairs. Prefer search_mentor_protege_agreements when you do not have a recordId. Example: {"recordId":"sba-mentor-protege:000e229d49161f817c51bc4df00f926a2424b37b70c0feaa662d7c03e6c3fb76"}
get_notice
ChatGPTFetch one source notice by recordId. Returns the most recently ingested typed source fields and provenance, without snapshots or lifecycle events; code labels add pscLabel, naicsLabels, setAsideCode and setAsideProgram, plus setAsideSamLabel when the source was a code. Use when you need raw notice fields for a row from list_notices or search_opportunities. Prefer get_opportunity when you need the summary, documents and analysis. Example: {"recordId":"02af3dac-b32a-5d91-9434-b6e1e7353703"}
get_nsn
ChatGPTFetch one current DLA PUB LOG catalog record by exact National Stock Number. Use when you already have a full NSN and need item identification and public release provenance. Returns {data} with NSN, NIIN, FSC, item name, source of supply, end item name and release hashes. No catalog-wide search, approved-supplier determination, pricing or award history. cancelled_niin is retained source text, not an active-status flag. Monthly replacement is manual with no delivery SLA. Prefer search_opportunities for solicitations and search_awards for awards. Example: {"nsn":"5310-01-234-5678"}
get_opportunity
ChatGPTFetch one SAM or DIBBS opportunity by recordId. Returns summary, documents, analysis and sourceUpdatedAt (when the row last changed, not an amendment date); code labels add pscLabel, naicsLabels, setAsideCode and setAsideProgram, plus setAsideSamLabel when the source was a code. Use when you need full detail for a row from search_opportunities or list_opportunity_analysis. Pass the same source the search used. SAM and DIBBS keep no ingestion history; SAM archive reads are retired. Prefer get_notice when you need raw source fields, or get_opportunity_documents when you only need the document block. Example: {"recordId":"7825b6f1-3c80-5641-b45a-1ca43b6aa4fd","source":"sam_current"}
get_opportunity_analysis_rollup
ChatGPTAggregate a bounded sample of opportunity analysis rows into facet buckets (counts by status, category, PSC, NAICS, set-aside and more) and feature totals (sampled rows with a PDF, an award block or technical documents). Returns buckets and totals; set-aside buckets add code and program. Use when you want a breakdown in one call. Takes the same filters as list_opportunity_analysis; limit sets the sample size (default 100, maximum 500). totalCount is the sampled row count. If countsAreCapped is true, more matching rows exist and every bucket, feature count and metric is sample-only, not an exact total. Prefer list_opportunity_analysis when you need the rows themselves. Example: {"source":"sam_current","naics":"5617","limit":100}
get_opportunity_document_text
ChatGPTFetch preserved document text for one opportunity. Returns plain text by default (format=text; omit documentId to pick a preserved result, or documentId plus page for one page) or, with format=structured and documentId, ordered page/block/cell fragments (limits: 20000 characters, 25 blocks, 100 cells, 16 MiB artifact). Use when reading a document or resolving a citation: pass representationId and artifactHash from evidence search plus page/block, then cursor; block and cursor need format=structured. PDF pages are physical; worksheets and Word units are labeled separately. References resolve only while the same bytes stay authorized and retained; changed or missing versions return 404, never remap. Unsupported covers return status only. Prefer search_opportunity_documents to find the passage first. Example: {"recordId":"7825b6f1-3c80-5641-b45a-1ca43b6aa4fd","maxChars":20000}
get_opportunity_documents
ChatGPTFetch the most recently ingested document associations and metadata for one opportunity: names, sizes, access state, collection status, hashes, source URLs and preserved parsedResults matched to the file bytes of the record. Returns the REST documents response unchanged. An attachment url is null when deleted, restricted or unavailable. Use when listing attachments before reading or searching their text. Pass the same source used by search_opportunities. SAM and DIBBS have no document version history; text and search use compatible preserved results with full document coverage. Prefer get_opportunity when summary or analysis is also needed, or search_opportunity_documents to find a phrase. Example: {"recordId":"7825b6f1-3c80-5641-b45a-1ca43b6aa4fd","source":"sam_current"}
get_opportunity_incumbents
ChatGPTEstimate likely incumbents for an opportunity from past award history (USAspending contract transactions dated 2022-04-04 or later). Each candidate has a confidence level and citations; candidates are estimates, not confirmed incumbents. Returns result=no_prior_award_found with the method used when nothing matches.
get_procurement_forecast
ChatGPTFetch one DHS APFS procurement forecast by recordId (dhs-apfs:<DHS numeric record ID>), as most recently ingested. Returns {data} with source dates, estimated value band, period of performance and business contacts; absent records return not-found. Use when you have a recordId from search_procurement_forecasts. Forecasts are revisable plans, not award commitments. Prefer get_opportunity when you need a published notice. Example: {"recordId":"dhs-apfs:67660"}
get_regulation
ChatGPTGet current captured metadata, abstract and official links for one procurement regulation by document_number. PDF/HTML bytes are not mirrored. Use search_regulations to find document numbers; coverage is 48 CFR plus selected-agency FAR documents, not all agencies. Example: {"document_number":"2026-12345"}
get_subaward
ChatGPTFetch one reported contract subaward by recordId. Returns {data} with reported work, exact decimal amount, dates, both parties, exact references, current/not_returned status and capture provenance; missing IDs return not-found. Use when you have a recordId from search_subawards. Not-returned records keep last-known facts and do not prove termination. Prefer search_subawards when you do not have a recordId. Example: {"recordId":"usaspending-contract-subawards:f7c65f5daeb07b20924d629e34a27be56a117a7764053a1cadbcec6126557952"}
get_vehicle_contract
ChatGPTFetch one GSA eLibrary contract by recordId. Returns {data} with all category/SIN memberships, category-specific purchasing flags and catalog links, supplier UEI, reported classifications, end dates and source/report dates; absent records return not-found. Use when you have a recordId from search_vehicle_contracts. These are vehicle holdings, not order amounts. Prefer search_vehicle_contracts when you do not have a recordId, get_entity for SAM registration, or search_awards for spending. Example: {"recordId":"gsa-elibrary:00243cf3f1260a43b7b8c4a5367de3e9ca913b53a3cea3e0e6086ed422377c0a"}
get_wage_determination
ChatGPTFetch one cited wage determination by exact WD number. Use when inspecting a known WD. Prefer search_wage_determinations when finding WD numbers. Returns {data} with provenance, nullable documentTextUrl and up to 100 distinct citing notices ordered by record ID; citingNoticeCount is the full count. Missing numbers return not found, not proof no official determination exists. Cited-only SCA/DBA determinations from held notices, not the national SAM.gov inventory. Highest official revision among held cited documents; description-only citations can have null document fields. Verify applicability and rates on SAM.gov before use. Example: {"wd_number":"2015-5227"}
list_award_rollups
ChatGPTList contract action totals grouped by agency, vendor UEI, NAICS or PSC (groupBy). Returns top-N groups by total obligated amount (default) or award count. Use when comparing top buyers or winners. Prefer search_awards when you need records. Unfiltered totals are precomputed (calculatedAt = snapshot). Fiscal-year-only rollups may come from a nightly precomputed view (with calculatedAt); other date ranges compute on demand and may return 503 query_timeout on large scopes. Amounts sum signed obligations; counts are distinct awards; contract value counts once. Filtered vendor rollups skip pre-UEI rows. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. limit defaults to 10 groups (max 25); no cursor. Example: {"groupBy":"awarding_agency_code","limit":10}
list_entity_exclusions
ChatGPTList SAM.gov exclusions for one UEI, as most recently ingested. Returns matching records (recordId, exclusion type/program, excluding agency, active/termination dates) and a cursor; an empty list means no exclusion row matches that UEI. Use when screening one known vendor. Prefer search_exclusions when you have a CAGE, NPI or name instead, or get_entity_profile when you also need awards and integrity records. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"uei":"QT8CKKV6LTB5"}
list_notices
ChatGPTList SAM and DIBBS source notice fields with cursor pagination. Returns one row per notice with source notice fields (noticeId, solicitation number, title, notice type, dates, PSC, NAICS; responseDeadlineDate is the source's deadline date); SAM returns the most recently ingested active record per notice, including older notices still active, and archive versions are retired. Use when you need raw source fields or want to walk every notice from one source. Prefer search_opportunities when you need filters or search. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"source":"dibbs","limit":10}
list_opportunity_analysis
ChatGPTList opportunities with structured filters (posted or closing window, agency, status, category, PSC, NAICS, set-aside, notice type, place of performance, office) and boolean feature flags (hasPdf, hasAward, hasExtractedText, hasTechnicalDocuments, hasDescriptionText). Returns enriched analysis rows with summary, dimensions and features. Use when the question maps to filters rather than free text. Pricing type and contract vehicle filters are unavailable. Prefer search_opportunities when searching free text or NSNs, or get_opportunity_analysis_rollup when you need counts. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"source":"sam_current","naics":"5617","hasPdf":true}
list_opportunity_related
ChatGPTList cross-source relationships for one opportunity. Returns matching DIBBS records for a SAM record (or SAM records for a DIBBS one) with relationType, confidence, direction and identifiers, including evidence for SAM/DIBBS links. Other SAM notices in the same solicitation are not listed: relatedRecordCount counts those, including the record itself, so it can exceed zero when this returns none. Use when the user asks for cross-source matches or whether a SAM posting points to DIBBS. Prefer search_opportunities when starting from keywords instead of a known recordId. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"recordId":"7825b6f1-3c80-5641-b45a-1ca43b6aa4fd","limit":10}
list_size_standards
ChatGPTList the published SBA size-standard table or look up one exact six-digit NAICS code. Use when checking thresholds and every exception for an industry, not a company's eligibility. Returns {data,page,notice} with full footnotes, sector conditions and workbook provenance; thresholds distinguish annual receipts, assets and employee counts. Unknown NAICS returns an empty page. No company inputs or historical table selection. Prefer search_entities for supplier registration facts; this tool does not determine whether a business is small. All row fields, including null conditions and full footnotes, are preserved. Keep the NAICS filter unchanged with page.nextCursor; changed workbook bytes invalidate old cursors (restart on 400). Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"naics":"541330","limit":10}
list_sources
ChatGPTList datasets with endpoints, ingestion timestamps, record counts and supported selectors. Returns {data, meta}, including additive metrics.recordCount, recordKind and latestSourceDate. Use when choosing a dataset or checking when one was last ingested before you search. Prefer the dataset's own search tool when you need records. Counts are retained canonical records (awards count actions, not award summaries) and do not establish coverage; null counts mean unverified, and a null lastCompletedRefreshAt means a finished import is unproven. A null snapshot parameter means history selection is unsupported. SAM exposes the most recently ingested opportunities only; archive and date-pinned history are retired. The combined catalog keeps its published expiry, capped at 26 hours, and fails closed when unavailable. Example: {"dataset":"awards"}
search_agencies
ChatGPTSearch SAM.gov Federal Hierarchy organizations, as most recently ingested. Returns candidates with fhOrgId, name, type, status, parent and agencyCode. Use when matching a department, independent agency, sub-tier, name, ID, or code, such as the agencyCode search_opportunities filters on. Name words may appear in any order; bounded typos are tolerated. parentFhOrgId lists children. Prefer get_agency when you already have the fhOrgId. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"veterans affairs","status":"ACTIVE","limit":5}
search_assistance_listings
ChatGPTSearch SAM.gov Assistance Listings (CFDA federal grant and assistance programs), as most recently ingested. Filter by listing id, organization codes, fiscal year, status, or text query. Returns program rows with Federal Hierarchy enrichment when organization codes match. Use when identifying a federal assistance program or its listing number. Prefer search_opportunities when looking for contract solicitations. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"departmentCode":"1200","status":"Active","limit":10}
search_awards
ChatGPTSearch USAspending contract and IDV transactions by keyword, vendor, agency, NAICS/PSC, PIID, place state, fiscal year or action dates. Returns rows (newest action_date first) with recordId and nullable sourceUpdatedAt. Use when finding individual awards. Keyword searches cover only transaction actions dated 2022-04-04 or later. Prefer list_award_rollups when you need group totals, or get_award for a known recordId. Vendor fields are null on rows dated before Apr 4, 2022 (D&B-sourced) or with unknown action dates; vendorFieldsWithheldReason identifies withholding, otherwise null. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"keyword":"janitorial","placeState":"TX","limit":10}
search_documents
ChatGPTSearch preserved SAM and DIBBS attachment text. Returns evidence snippets with opportunity context and pagination. Use when the words are inside documents: FAR/DFARS clauses, NSNs, part numbers or phrases. Prefer search_opportunities when searching titles or metadata, or search_opportunity_documents when the opportunity is known. Results include closed solicitations unless activeOnly=true. SAM reads only results with full document coverage matching the public file hashes of each matched record; reads never trigger parsing. Ranking caps at 1,000 passages: selective queries choose the best scores, then active opportunities, then newest posted dates before the cap; very broad or unprobed queries take the first 1,000 in stable content/parser/chunk order, then rank that set. If ranking.truncated is true, narrow the query or scope by recordId/threadKey. A file shared by notices of one thread appears once, others in alsoAttachedTo. evidence=true adds page/block pins (evidenceStatus='aligned') for get_opportunity_document_text format=structured; unaligned results return evidence=null and evidenceStatus='alignment_pending'. Beta: broad phrases may be slow. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"CMMC","activeOnly":true,"limit":10}
search_entities
ChatGPTSearch SAM.gov registrations by UEI, CAGE, name, status or primary NAICS, physical US state/territory and registration expiration. Returns registrations with dates, city/state and applied-filter meta. Use when finding a vendor's UEI or filtering registration status and expiry. query matches every word in any order. Prefer get_entity_profile when you need one UEI's business types, certifications, awards or risk records. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"veteran janitorial","registrationStatus":"ACTIVE","state":"VA","limit":10}
search_exclusions
ChatGPTSearch active SAM.gov exclusions (debarments, suspensions and other ineligibility) by UEI, CAGE, NPI or text, as most recently ingested. Returns rows with recordId, name, exclusion type/program, excluding agency and dates. Use when screening a person or firm for debarment or suspension. Prefer list_entity_exclusions when you have exactly one UEI, or get_exclusion for one recordId. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"saint petersburg exchange","limit":10}
search_integrity_records
ChatGPTSearch SAM.gov Responsibility/Qualification and FAPIIS integrity records by UEI, CAGE, family, extract kind or text, from the most recently ingested report pair. Returns records with recordId, firm, family, date, agency, title and disposition. Use when checking a vendor's terminations, proceedings or responsibility history. Active status is unavailable (statusActive is unsupported). Prefer get_entity_profile when you also need exclusions and awards for one UEI. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"recordFamily":"termination","limit":10}
search_labor_rates
ChatGPTSearch GSA CALC+ MAS contractor labor rates: fully burdened hourly ceiling rates by labor category. Use when asked for a ceiling rate, labor rate, hourly rate or pricing for a role (e.g. senior security engineer), for price-to-win, or to benchmark a supplier. Returns {data,page} with exact USD/hour decimal strings for three contract years, qualifications, SIN labels, provenance and nullable exact eLibrary links. Ceilings, not prices paid or salaries. Prefer search_vehicle_contracts when researching memberships, or search_awards for spending. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"senior security engineer","maxPrice":"250","limit":10}
search_mentor_protege_agreements
ChatGPTSearch the SBA active mentor-protégé list, as most recently ingested. Returns {data,page} with both firms, NAICS, states, approval date, report/capture dates, raw and nullable validated identifiers, status and nullable exact company references. Use when finding approved pairs by firm UEI/name, role, NAICS, state or approval date. Certification text is source-reported, not an eligibility decision; a pair does not establish a JV, ownership or subcontract. Prefer search_entities when you need registration facts, or search_awards for spending. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"naics":"541330","mentorState":"MA","limit":10}
search_opportunities
ChatGPTSearch SAM and DIBBS opportunities by free text, dates, agency, PSC, NAICS, set-aside, notice type, place or NSN. Returns summaries with recordId and nullable row-level sourceUpdatedAt (last row change, not an amendment date). Use when finding solicitations to bid on. DIBBS rows carry summary.nsns (13-digit NSNs) and summary.partNumbers; pass source=dibbs with nsn for an exact NSN. Use sortBy=closing for upcoming deadlines (from now on, soonest first). Prefer list_opportunity_analysis when filtering on document or award flags. If empty, meta.emptyResult names the filter and up to 3 fixes. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields plus the decoded code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"query":"janitorial","setAside":"SDVOSB"}
search_opportunity_documents
ChatGPTSearch preserved attachment text for one known SAM or DIBBS opportunity. Returns evidence snippets with character offsets and pagination. Use when checking what one solicitation's documents say about a term. Results include closed solicitations unless activeOnly=true (closing date in the future). SAM matches results with full document coverage to the public file hashes of this record without reparsing. Set evidence=true for exact artifact and page/block ranges (evidenceStatus='aligned'); unaligned matches return evidence=null with evidenceStatus='alignment_pending'. Read aligned pins with get_opportunity_document_text format=structured. Prefer search_documents when the opportunity is unknown. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"recordId":"7825b6f1-3c80-5641-b45a-1ca43b6aa4fd","query":"wage determination","evidence":true}
search_procurement_forecasts
ChatGPTSearch DHS APFS procurement forecasts: planned acquisitions published before a solicitation, as most recently ingested. Returns {data, page} with component, NAICS, status, estimated timing/value bands and source links. Use when seeking early leads before a notice is posted. Plans can change or be canceled; dates and amounts are estimates, and no-longer-required records stay listed. Prefer search_opportunities when you need published notices, or get_procurement_forecast for contacts. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"naics":"541330","contractStatus":"REC","limit":10}
search_regulations
ChatGPTSearch Federal Register procurement regulation metadata, abstracts and official links: 48 CFR plus selected-agency FAR documents, not the all-agency Federal Register. PDF/HTML bytes are not mirrored. Use for procurement rule research; use search_opportunities for solicitations. Inclusive publication dates; newest publication_date and document_number first. Current rows may change between pages; keep filters unchanged with cursor. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"cfr_title":48,"q":"acquisition","limit":10}
search_subawards
ChatGPTSearch first-tier contract subawards reported from October 1, 2024 forward. Returns {data,page} with exact USD decimal strings, both reported parties, work description, source/report/capture metadata and nullable exact company/prime-award references. Use when finding which subcontractors a prime reports using, or which primes a firm works under. reportedStatus defaults to current. Values are reported, not verified payments; grants are excluded. Prefer search_awards when researching prime spending. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"primeUei":"VM45UTL9JTH1","limit":10}
search_vehicle_contracts
ChatGPTSearch the public GSA eLibrary contract directory, as most recently ingested (excludes BMOS1, BMOU1, HCATS, HCATS 8A, OASIS and USACON). Returns {data,page} with contract facts, compact category codes and source/report dates. Use when finding suppliers on a vehicle or category/SIN, or the vehicles one UEI holds. Reported indicators are not order-level eligibility. Prefer get_vehicle_contract when you need all memberships, or search_awards for spending. Rows are compact by default (no empty fields, raw duplicates or bookkeeping; long text cut, see truncated); verbose: true returns the full REST fields; this tool adds no code labels. Pages default to 10 rows (max 25); continue with page.nextCursor. Example: {"vehicle":"MAS","category":"541611","limit":10}
search_wage_determinations
ChatGPTSearch cited wage determinations by act, state, WD number or citing opportunity. Use when finding labor determinations cited in federal notices. Returns {data,page} with full provenance and nullable document fields. Prefer get_wage_determination for one WD's citing notices and held document-text link; prefer search_labor_rates for GSA contract labor ceilings, not statutory wages. Cited-only SCA/DBA determinations from held notices, not the national SAM.gov inventory. Highest official revision among held cited documents; description-only citations can have null document fields. Verify applicability and rates on SAM.gov before use. Pages default to 25 rows (max 200); API plan limits apply. Continue with page.nextCursor. Keep filters unchanged when following page.nextCursor; accepted-data changes invalidate cursors. Example: {"act":"SCA","limit":10}
get_agency
Claudeget_assistance_listing
Claudeget_award
Claudeget_award_hierarchy
Claudeget_entity
Claudeget_entity_profile
Claudeget_exclusion
Claudeget_integrity_record
Claudeget_labor_rate
Claudeget_mentor_protege_agreement
Claudeget_notice
Claudeget_nsn
Claudeget_opportunity
Claudeget_opportunity_analysis_rollup
Claudeget_opportunity_document_text
Claudeget_opportunity_documents
Claudeget_opportunity_incumbents
Claudeget_procurement_forecast
Claudeget_regulation
Claudeget_subaward
Claudeget_vehicle_contract
Claudeget_wage_determination
Claudelist_award_rollups
Claudelist_entity_exclusions
Claudelist_notices
Claudelist_opportunity_analysis
Claudelist_opportunity_related
Claudelist_size_standards
Claudelist_sources
Claudesearch_agencies
Claudesearch_assistance_listings
Claudesearch_awards
Claudesearch_documents
Claudesearch_entities
Claudesearch_exclusions
Claudesearch_integrity_records
Claudesearch_labor_rates
Claudesearch_mentor_protege_agreements
Claudesearch_opportunities
Claudesearch_opportunity_documents
Claudesearch_procurement_forecasts
Claudesearch_regulations
Claudesearch_subawards
Claudesearch_vehicle_contracts
Claudesearch_wage_determinations
Claude