MCP App Store
by KeenEthics
Ex Parte icon

Ex Parte MCP Server for Claude and ChatGPT

By ExParte Inc

Overview

Bring Ex Parte research into Claude with your Enterprise account. Find U.S. patent-litigation cases, patents and reexaminations; explore parties, attorneys, law firms and judges; and inspect case dockets. Get source documents, read passages with page references, and inspect supported scanned-page images. Open original documents through Ex Parte links, review existing analytical reports alongside the filings, and continue longer source text in sections. Interactive case and firm cards keep records and sources together. Optional bookmark permission lets you save useful items. MCP must be enabled for your account, and record permissions and document availability apply. The connector reads existing reports; it does not generate new reports or automatically update court dockets.

How to connect

  1. Claude: Open the Ex Parte listing in the Claude connectors directory and choose Connect. Claude listing Ex Parte exposes 34 tools here.
  2. ChatGPT: Open the Ex Parte listing in the ChatGPT app directory and choose Connect. ChatGPT listing Ex Parte exposes 34 tools here.

Tools

bookmark_status

ChatGPT
Returns whether one supported case, attorney, law firm, party, patent or reexamination is saved in the current account. The result is a saved-status Boolean for the exact identified item. This operation does not add or remove a bookmark.

download_document

ChatGPT
Gets one identified dct: or cafc: docket document for an eligible Enterprise account, returning page-marked text (up to 24,000 characters) and its Ex Parte link. A multi-document CAFC filing may contain only the first document. Availability and account permissions apply. Retrieval can make the requested document available in Ex Parte. Ask for missing case or docket identifiers instead of selecting a document from previous chats, recent activity or bookmarks.

find

ChatGPT
Looks up cases by number or caption, patents by number, attorneys by name, and ex parte reexaminations by control number, patent number or title words. Results contain typed record references and up to limit matches per selected type. The types field restricts the record categories; reexamination title lookup requires that category explicitly. Default categories depend on the shape of the query. Reexamination results require the corresponding account permission. Empty results mean no match was returned for this lookup.

get_attorney

ChatGPT
Returns sections for one identified attorney, a portal link and an interactive card in compatible clients. Defaults: available identity, firm, experience, contact, practice areas and a bounded recent-case preview. Optional court-scoped ratings and trends. Follow structured metricDefinitions for the practice-area denominator, combined share, completeness and rounding. A shorter Markdown preview never renormalizes shares or establishes the card's visible row count. Incomplete source counts have no percentage. Profile and court aggregates do not prove distinct cases, matched cohorts or the cause of differences from overlapping area counts. Rating totals and case-list counts can describe different populations.

get_case

ChatGPT
Returns selected portal sections for one identified District Court, PTAB, Federal Circuit or ITC proceeding. Default sections include the overview and existing reports; other supported sections cover participants, patents, related proceedings, key events and court-specific records. Results include a portal link, section freshness and available navigation metadata; overview results also provide an interactive case card in compatible clients. District Court and Federal Circuit data are stored records; PTAB and ITC use snapshots. No court update or new report is initiated.

get_firm

ChatGPT
Returns selected profile sections for one identified law firm, with a portal link and an interactive profile card in compatible clients. Default content includes firm identity, bounded client and opponent analytics, top attorneys, recent-case previews and the PTAB institution-denial panel. Additional sections expose ratings, trends, practice areas and analytics. Each analytics row retains its court and population scope; case and appeal counts can overlap and do not establish distinct clients or direct client relationships.

get_judge

ChatGPT
Returns selected portal profile sections for one District Court or Federal Circuit judge identified by a court-qualified reference. Default content includes the overview, available party, counsel and technology-center previews, recent cases and cases by year. Optional sections include ratings, analytics, outcomes and assignments by year. Results provide a portal link and an interactive profile card in compatible clients. Court scope is part of the judge identity.

get_party

ChatGPT
Returns selected portal profile sections for one party reference, with a portal link and an interactive profile card in compatible clients. Default content includes the overview, available counsel, opponents and technology-center previews, plus a bounded recent-case preview. Optional sections expose related entities, court-scoped analytics, ratings and institution-denial data. The party name is matched ignoring case and surrounding punctuation, and & equals and: a search result that differs from the user's wording only in punctuation is the same company. When no portal party matches, the result is a not_found error with up to five candidate refs and no profile.

get_patent

ChatGPT
Returns selected sections for one US patent, including available identity, title, assignee, examiner, art unit, dates, expiration, abstract, CPC classifications, analytics, reexaminations and stored-PDF availability. Optional report-status sections describe the current organization's existing prosecution report and any ongoing generation status. This operation does not start a report or establish that the patent PDF has been read.

get_ptab_institution_statistics

ChatGPT
Returns PTAB institution-decision statistics since 2021 for supported IPR, PGR, CBM and DER proceedings. An optional exact law-firm or party subject scopes results to its selected side; without subject the measure covers all eligible parties under the filters. The optional baseline uses the same filters. Measures include institution outcomes, Director or panel decisions, and doctrines in discretionary denials, with supported breakdowns. Denial grounds and decision dates retain their stated scope; comparisons do not establish causation or the quality of legal representation.

get_reexam

ChatGPT
Returns selected sections of one ex parte reexamination from the stored nightly pipeline snapshot, including available patent identity, status, claims, parties, related cases, file-wrapper document metadata and transaction history. The corresponding account permission is required. This operation does not trigger a USPTO refresh or establish that listed documents were read.

get_reexam_statistics

ChatGPT
Returns counts, rates or pendency measures for US ex parte reexamination requests filed since 2020. An optional exact law-firm or party subject scopes the measure to its selected side; without subject the measure covers all eligible requests under the filters. The optional baseline uses the same filters. Measures cover filings, order outcomes, claim outcomes and pendency, with supported time or entity breakdowns. Results describe their stated cohort and denominators; they do not establish causation or the quality of legal representation. Reexamination permission is required.

list_bookmarks

ChatGPT
Returns the current account's saved-item counts by supported category, or a searchable and sortable page from a selected saved-item list. Categories include cases, attorneys, law firms, parties, patents and reexaminations. Rows contain typed references where portal identities are available. This operation reads saved items without changing them.

list_case_dockets

ChatGPT
Returns stored docket entries for one identified case, newest first, with available document metadata. Results contain up to 100 entries per page, subject to response-size limits, and continuation metadata when more entries remain. Optional filters select an exact docket number or key events. A docket entry or document-availability flag does not establish that its PDF text was returned. This operation does not update court records.

list_entity_analytics

ChatGPT
Returns one paginated analytics table for a firm, party, judge or court, such as clients, opponents, counsel, technology centers or litigated patents. Available tables and filters depend on entity and court. Counts belong to their returned population and may overlap. In litigated-patent tables, side describes litigation role rather than patent ownership; assignee identifies the owner. A patent encountered on both sides remains one row with multiple sides.

list_entity_cases

ChatGPT
Returns a paginated, filterable case list for one attorney, law firm, party, judge or District Court in a selected court system. Supported filters and sorting vary by entity and court; unsupported combinations return the accepted options. Rows include typed case references. District Court judge lists can include past assignments; firstJudgeEntry, lastJudgeEntry and currentJudge describe the assignment relationship separately from the case filing date and current court.

list_itc_attachments

ChatGPT
Lists available public EDIS attachment metadata for one ITC docket entry or known attachment list. Results include attachmentListId and attachmentId for exact source identification. The account's ITC document-download permission is required. Listing an attachment does not return its source text or establish successful PDF delivery.

list_law_firm_attorneys

ChatGPT
Returns a paginated list of attorneys associated with one law firm in a selected court system. Supported search and sort fields include the available experience and rating measures. Each returned attorney has a typed reference; values describe the selected court scope.

list_patent_records

ChatGPT
Returns a paginated record collection for one patent: linked litigation proceedings, forward citations, backward citations, non-patent citations or assignment history. Linked proceedings include typed case references. Published-application citations are marked as applications and identify an applicant rather than an assignee. Continuation metadata describes remaining records in the selected collection.

list_recent_cases

ChatGPT
Returns a paginated list of cases the user recently opened in the Ex Parte portal, with available typed case references and portal links. The list represents recorded portal views. Assistant reads do not create new entries in that view history.

list_reports

ChatGPT
Lists existing analytical reports and report-status records for an identified case or patent. Supported collections include District Court complaint reports, PTAB petition and docket reports, related-patent PDF availability, the current organization's prosecution-report statuses and its existing email templates. Availability depends on the requested collection and account permissions. This operation returns an inventory or status, does not generate a report, and does not by itself return the underlying filing text.

load_context

ChatGPT
Prepares a temporary research context for one identified case or patent, including a Markdown index, source identities, coverage information and up to two available PDFs. District Court contexts select the portal's primary complaint (the highest complaint docket with an existing analysis, else the highest complaint docket) with its matching existing analysis and name every newer complaint filing; PTAB contexts select the petition and matching analysis; patent contexts include the patent PDF and related cases. ITC and Federal Circuit contexts provide an index. Preparation creates temporary working material, does not generate analysis, and does not establish that the client received or read a PDF.

read_context

ChatGPT
Returns a character-range excerpt of a prepared context index and, optionally, its available PDF resources or links. The result includes the current position, continuation and index-completion information. Index completion is separate from document-inventory completeness and PDF reading coverage. The context belongs to the connection that prepared it and expires after one hour; temporary download links have their own expiry.

read_context_file_page

ChatGPT
Returns one selected page of a prepared context PDF as a bounded JPEG, with source identity, SHA-256, page number, total pages and continuation metadata. Suitable for scanned text, figures and drawings in clients that support image results. Page numbers are one-based. Current access and the prepared file hash are checked on each request. A links-only result contains no page image; delivery of one page does not establish coverage of other pages. No new analysis or AI-generated media is produced.

read_context_file_text

ChatGPT
Returns page-marked extracted text from one PDF in a prepared context. Inputs identify the context, file and character range. Each request checks current access and the prepared file hash. Results include source identity, delivered range, continuation and extraction-coverage fields. Scanned or unreadable pages and extraction limits may leave source content unavailable; an end-of-text position alone does not establish complete PDF coverage. The context expires after one hour. Text delivery does not establish a native PDF attachment.

read_document

ChatGPT
Returns page-marked extracted text, source metadata and available PDF-resource or portal links for an identified document. Supported kinds are dctDocket, cafcDocket, dctExhibit, ptabDocket, itcAttachment, patent, patentProsecution and patentCitation, with the corresponding case, docket, attachment, patent or citation identity. District Court and Federal Circuit reads use stored sources; a public PTAB PDF may be obtained from USPTO. Results expose continuation and extraction coverage. Scanned, encrypted, invalid or oversized PDFs can produce links-only availability results; terminal text offsets alone do not establish complete PDF coverage. No native client attachment is guaranteed.

read_report

ChatGPT
Returns a character-range excerpt of an existing AI report, with its available report identity, source references and continuation metadata. Case-based kinds are complaint, ptabDocket and complaintSummaryLine; they require a case and docket number. Patent-based kinds are patentProsecution, patentExaminerSearch, patentInvalidity and patentEmail, with the applicable patent and template identity. Organization-scoped reports require a successful existing run from the current organization. Text chunks contain up to 8,000 characters. AI analysis is distinct from the original filing; this operation does not generate a new report.

resolve_cases

ChatGPT
Resolves case numbers, supported portal database IDs or caption words into typed case references for a selected court system: District Court, PTAB, Federal Circuit or ITC. Number and ID batches accept up to 25 entries. A District Court number can match more than one court; returned matches retain their court identities. Portal-ID and caption lookup support varies by court system.

search_attorneys

ChatGPT
Returns a paginated attorney list filtered by supported name, law-firm, client, opponent, technology, court-experience, rating and trend criteria. Rows include attorney references and available law-firm references. Entity-name filters require exact portal vocabulary values. Availability of experience and rating fields depends on the selected court scope.

search_cases

ChatGPT
Returns a paginated case list using portal search. Without caseType, query performs simple search across courts. With caseType, supported advanced filters apply within one court system, including case number, caption, participants, patents, dates, outcomes and District Court key events. At least one applicable filter is required. Results contain typed case references and the available pagination information; supported filters and sorting depend on the selected court system.

search_filter_values

ChatGPT
Returns portal vocabulary values and supported entity references for a selected kind, such as parties, law firms, judges, attorneys, patents, examiners or key events. Name fragments belong in name. Court-specific kinds require caseType; attorney and patent facet kinds have separate scopes. REEXAM supports law firms and parties with reexamination permission. Returned judge vocabulary may include former and magistrate judges and is not a roster of currently sitting judges.

search_patents

ChatGPT
Returns patent records matching supported number, title, assignee, examiner, technology-center, agent, art-unit and issue-date filters. A number alone performs number lookup; advanced filters can be combined with a number. The numbers field checks a batch of up to 25 patent numbers for presence in the portal. Results include available patent identity fields and pagination information.

set_bookmark

ChatGPT
Sets whether one supported item is saved in the current account. saved=true adds the bookmark; saved=false removes it. This changes account workspace data and requires portal:write permission. Repeating the same requested saved state is idempotent. The target identifies one case, attorney, law firm, party, patent or reexamination.

whoami

ChatGPT
Returns the Ex Parte account associated with the current connection, including its available role, permissions and connection scopes. The optional includeProfile field adds directory profile details such as job title, company and phone, subject to access. This operation does not change account access or the subscription.

bookmark_status

Claude

download_document

Claude

find

Claude

get_attorney

Claude

get_case

Claude

get_firm

Claude

get_judge

Claude

get_party

Claude

get_patent

Claude

get_ptab_institution_statistics

Claude

get_reexam

Claude

get_reexam_statistics

Claude

list_bookmarks

Claude

list_case_dockets

Claude

list_entity_analytics

Claude

list_entity_cases

Claude

list_itc_attachments

Claude

list_law_firm_attorneys

Claude

list_patent_records

Claude

list_recent_cases

Claude

list_reports

Claude

load_context

Claude

read_context

Claude

read_context_file_page

Claude

read_context_file_text

Claude

read_document

Claude

read_report

Claude

resolve_cases

Claude

search_attorneys

Claude

search_cases

Claude

search_filter_values

Claude

search_patents

Claude

set_bookmark

Claude

whoami

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/ex-parte"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/ex-parte)