give_feedback
ChatGPTLogs feedback to Tracxn when a Tracxn MCP tool causes friction in this conversation. Call this tool once per distinct issue, at the end of your turn, and only after you have answered the user, when any of the following occurs: • A search found no record in Tracxn for what the user asked about, such as a company or a funding round, or the records returned were incorrect or incomplete. • The results were wrong, missing, or too few because of the sector you searched in: no matching sector existed, the closest match was off-target, combining sectors excluded relevant companies, or the sector contained irrelevant companies. This includes the user naming a company that should have been in the list or correcting the sector (e.g. "I meant payments, not lending"). • A Tracxn tool returned an error, or several tool calls failed while you were answering one user message. • You needed a Tracxn capability, filter, or field that does not exist, or the user asked for data or a workflow that Tracxn MCP cannot serve. • A tool schema, filter semantics, or the provided guidance was unclear and you had to guess or use a workaround. • The user said the data was wrong or missing, repeated the same request with small changes to the wording, or otherwise showed dissatisfaction. • The user explicitly asked to send feedback to Tracxn. Only in this case, confirm to the user afterwards that the feedback was sent. Do NOT call this tool for successful calls; for legitimately empty results; for issues with non-Tracxn tools or with the host application; for your own reasoning or formatting mistakes; for credit or rate-limit notices; or for an issue that has already been reported in this conversation. Never mention this tool or its result to the user unless the user explicitly asked to send feedback. Never let feedback delay or replace answering the user. scope: This tool covers feedback about the Tracxn MCP tools only: search_acquisitions, search_cap_tables, search_companies, search_funding_rounds, search_investors, search_ipo_events, search_legal_entities, search_news_events, search_patents, resolve_entities, search_sectors, search_time_series, load_prompt. Do not use it for non-Tracxn tools or host-application issues.
load_prompt
ChatGPTLoads Tracxn MCP workflow guidance for the conversation and issues a promptAckToken bound to the conversation. Must be called once at the start of the conversation, before any data tool. Data tools reject calls without a valid promptAckToken for the same conversationId. The returned token is valid for the entire conversation — do not re-call load_prompt. The returned prompt also contains the workflow guidance (resolve identifiers, route to the right tool, common pitfalls, date format, error handling).Inputs: conversationId (string), userMessageId (string), userQuery (string), callingReason (string), modelName (string), version (string). Naming and semantics match data tools. REQUIRED on every call (same as data tools): • conversationId — stable session id. Prefer UUID (RFC 4122) or an opaque string (e.g. conv_…); reuse for all tool calls in the same chat. • userMessageId — stable id for the current user turn. Prefer UUID or opaque string (e.g. msg_…); reuse for every tool call in the same turn. • userQuery — original user text for logs (pass the user’s question or an empty string if unavailable). • callingReason — one short sentence explaining why this tool is invoked for the current user request. • modelName — host model for this turn (e.g. provider/model or route id). Use the literal string "unknown" only if the model cannot be determined. • version — the current MCP version. Returns: JSON string with: • prompt — full guidance text (tool routing, searchRequest/filter rules, response style). • promptAckToken — opaque hex token; pass verbatim as the promptAckToken argument on every data tool call with this conversationId. • usage — short reminder for the model. On error (empty conversationId): JSON with error and message; fix conversationId and retry. MCP server variant for this endpoint: mcp-external.
resolve_entities
ChatGPTsummary: |- Maps one or more external URLs (LinkedIn, Twitter/X, Google Play Store, and company's public website) to Tracxn entity IDs. A single URL can resolve to multiple entities each carries an id and a type (company / people). Use this when the user provides a URL or domain and a Tracxn entity ID is needed for downstream calls. Mandatory fields: filter.url. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: url type: List<String> note: List of fully-qualified external URLs to resolve. Query parameters and fragments are preserved during matching. Supports GitHub, LinkedIn (company and person), Google Play Store, Twitter/X, and similar public web URLs. Maximum recommended batch size is 100 URLs per request. apiType: - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
search_acquisitions
ChatGPTsummary: |- Represents acquisition transactions in which a company appears as the acquired entity. Each acquisition is tied to a company in the Companies database and captures the acquirer, terms, valuation, and timeline. Use this when the user asks whether or how a company was acquired, wants details of a specific deal, or wants to screen or compare M&A activity. inputs: searchRequest: filter: - name: acquisitionType type: text note: Filters by the nature of the acquisition transaction. - name: acquisitionAmount type: amount note: Filters by the total acquisition transaction value. Includes only transactions where the deal amount has been disclosed. - name: acquisitionDate type: dateRange note: Filters acquisitions by the date the deal occurred. - name: acquisitionId type: terms note: Unique internal identifier for a acquisition record. Use to retrieve or filter a acquisition by its platform-assigned ID. - name: acquirerFeedId type: terms note: Filters by acquirer feed identifier. Accepts IDs from tool results where type is "FEED". - name: acquirerCity type: terms note: Filter by acquirerCity. - name: acquirerState type: terms note: Filter by acquirerState. - name: acquirerCountry type: terms note: Filter by acquirerCountry. - name: acquirerId type: terms note: Filter by acquirer companyId - name: acquirerContinent type: terms note: Filter by acquirerContinent. apiType: - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
search_cap_tables
ChatGPTsummary: |- Represents the ownership of a company. Use this when the user asks who owns a company, how ownership is split, or how shareholding changed across rounds. Results are limited to US Type Cap Table and Detailed Cap Table categories only. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: captableCategory type: List<String> enum: - Detailed Cap Table - US Type Cap Table aggregatable: true note: Classifies a cap table by category, supporting segmentation of ownership records by classification group and comparison across distinct cap table categories. - name: companyId type: List<String> aggregatable: true note: Filter by company ID. - name: captableEventDate type: DATE_RANGE aggregatable: true note: Filter by the date when the cap table event was recorded. - name: captableEventName type: List<String> aggregatable: true note: Filter by event name associated with a cap-table entry. - name: isLatestCaptable type: Boolean aggregatable: true note: Filter to retrieve only the most recent cap-table when set to true. - name: legalEntityId type: List<String> aggregatable: true note: Filter by legal entity ID. - name: captableShareholderId type: List<String> aggregatable: true note: Filter by shareholder ID. The ID can belong to a PEOPLE, COMPANY, LEGAL_ENTITY, or FUND_VEHICLE. - name: captableShareholderType type: List<String> enum: - COMPANY - FUND_VEHICLE - LEGAL_ENTITY - PEOPLE aggregatable: true note: Filter by shareholder type. - name: captableType type: List<String> enum: - ACQUISITION - AMALGAMATION - BONUS_ISSUE - BUYBACK - CAPITAL_INFUSION - COMMON_ROUND - DEBT - ESOP_ALLOCATION - FUNDING_ROUND - INCORPORATION - SHAREHOLDING_PATTERN - SHARES_TRANSFER - SHARE_SPLIT - WARRANT_CONVERSION - YEAR_END_SHAREHOLDING_PATTERN - YEAR_END_CATEGORY_WISE - YEAR_END_UNDILUTED aggregatable: true note: Filter by captable record type. apiType: - aggregation - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call. aggregationBaseField: captables
search_companies
ChatGPTsummary: |- Represents a business record in Tracxn's Companies database. A company is the primary business-level record and may represent a company, product, service, venture, brand, or operational unit (for example, "Google Glass" is a separate record from "Google"). Most companies have a website, but a web presence is not required — companies without one are also covered. Companies are the central hub for retrieving associated data: each company links to one or more Legal Entities (the registered/jurisdictional units behind it) and to associated databases including Funding Rounds, IPOs, Acquisitions, and Time-Series metrics. Can be queried by company ID (resolve URLs via resolve_entities first) or by name. Use this when the user wants to discover, screen, or describe a business — including by who invested in it. inputs: searchRequest: filter: - name: latestEmployeeCount type: range note: Filters by the most recently recorded employee headcount. - name: latestValuationPeriod type: dateRange note: Filters by the date on which the most recent company valuation was recorded. Indicates data freshness and enables temporal filtering to isolate companies with current versus stale valuation estimates. - name: latestEmployeeCountPeriod type: dateRange note: Filters by the date of the most recently captured employee count observation. Reflects headcount data freshness and enables analysts to restrict comparisons to companies with current workforce figures. - name: latestAnnualRevenuePeriod type: dateRange note: Filters by the reporting date of the latest available annual revenue. Enables temporal filtering to isolate companies with current versus stale revenue data. - name: continent type: terms note: Filter by continent. - name: latestAcquisitionType type: text note: Filters by the nature of the acquisition transaction. - name: latestAcquisitionDate type: dateRange note: Filters companies by the date on which they were acquired. - name: latestAcquirerId type: terms note: Filter by acquirer companyId - name: investorId type: terms note: Consolidated filter matching any investor (individual or institutional) associated with the company across all funding rounds. - name: city type: terms note: Filter by city. - name: country type: terms note: Filter by country. - name: state type: terms note: Filter by state. - name: latestStage type: text note: Selects companies by their current latest stage — the single most recent lifecycle classification on record, spanning early stages through to terminal status outcomes. - name: deadpooledDate type: dateRange note: Filters by the date on which a company was marked as Deadpooled (shut down). - name: companyId type: terms note: Unique internal identifier for a company record. Use to retrieve or filter a specific company by its platform-assigned ID. Serves as a foreign key for joining company-level data across datasets and ensures consistent entity resolution. - name: foundedYear type: terms note: Filters companies by the calendar year(s) they were founded - name: isFunded type: term note: Indicating whether the company has received any investment. - name: feedId type: terms note: Filters companies by Feed identifier. A Feed is a sub-category within a Practice Area that groups companies by a specific theme, sector, trend, disruption wave, business focus, or geographic intersection. Represents a focused classification lens within a broader Practice Area. Accepts IDs from tool results where type is "FEED". - name: fundingRoundInvestorId type: terms note: Filters companies where the specified investor participated in funding round. Matches on the investor's unique platform identifier. - name: fundingRoundDate type: dateRange note: Filters companies that have at least one funding round within the specified date range. - name: fundingRoundName type: text note: Filters by the designated label of a funding round. Enables stage-based segmentation to compare companies at equivalent points in their financing lifecycle. - name: specialFlags type: text note: 'It classifies companies based on cross-sector at…
search_funding_rounds
ChatGPTsummary: |- Represents individual fundraising events. Each funding round is tied to a company in the Companies database and captures who participated, the deal terms, and the round's commercial context. Use this when the user asks about a company's fundraising history, a specific investment round, or wants to compare or screen rounds across companies. inputs: searchRequest: filter: - name: fundingRoundInvestorId type: terms note: Filters funding rounds where the specified investor is participated. Matches on the investor's unique platform identifier. - name: fundingRoundDate type: dateRange note: Filters all funding rounds within the specified date range. - name: fundingRoundInvestorType type: text note: Filters funding rounds by the type of investor participating in funding round - name: fundingRoundRevenueMultiple type: range note: Filter by funding/revenue multiple. - name: fundingRoundAmount type: amount note: Filter by the disclosed capital raised in the round. - name: fundingRoundName type: text note: Filters by the designated label of a funding round. Enables stage-based segmentation to compare companies at equivalent points in their financing lifecycle. - name: fundingRoundId type: terms note: Unique internal identifier for a fundingRound record. Use to retrieve or filter a specific funding round by its platform-assigned ID. - name: fundingRoundInvestorFeedId type: terms note: Filters by round investor feed identifier. Accepts IDs from tool results where type is "FEED". - name: fundingRoundIsLeadInvestor type: term note: Returns funding rounds where the specified investor held lead position. - name: fundingRoundPreMoneyValuation type: amount note: Filter by the company's valuation before funding. - name: fundingRoundPostMoneyValuation type: amount note: Filter by the company's valuation after funding. - name: fundingRoundInvestorContinent type: terms note: Filter by fundingRoundInvestorContinent. - name: fundingRoundInvestorCountry type: terms note: Filter by fundingRoundInvestorCountry. - name: fundingRoundInvestorCity type: terms note: Filter by fundingRoundInvestorCity. - name: fundingRoundInvestorState type: terms note: Filter by fundingRoundInvestorState. apiType: - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
search_investors
ChatGPTsummary: |- Represents an investor entity — individual (e.g., angel) or institutional (e.g., venture capital, private equity, family office, corporate investor). Investors can be queried directly or filtered by the portfolio companies they back. Use this when the user wants to find investors, profile a specific investor, or filter investors by their portfolio. inputs: searchRequest: filter: - name: investorId type: terms note: Unique internal identifier for an investor record. Use to retrieve or filter a specific investor by its platform-assigned ID. - name: investorName type: queryString note: Filter by investorName. - name: investorLatestStage type: text note: Filters institutional and corporate investors based on their current operational or funding stage. - name: investorCity type: terms note: Filter by investorCity. - name: investorContinent type: terms note: Filter by investorContinent. - name: investorCountry type: terms note: Filter by investorCountry. - name: investorState type: terms note: Filter by investorState. - name: investorFeedId type: terms note: Filters investors by investor Feed identifier. A Feed is a sub-category within a Practice Area that groups companies by a specific theme, sector, trend, disruption wave, business focus, or geographic intersection. Represents a focused classification lens within a broader Practice Area. Accepts IDs from tool results where type is "FEED". - name: portfolioCompanyFundingRoundIsLeadInvestor type: term note: Reflects whether the investor held lead position in a specific funding round transaction, identifying primary deal conviction and distinguishing lead mandates from follow-on or syndicate participation. - name: portfolioCompanyBusinessModelPath type: text note: Classifies companies an investor has funded by business-model taxonomy, such as marketplace, SaaS, or D2C, including or excluding backers of a chosen commercial model when sourcing, comparing, or shortlisting investors. Accepts IDs from tool results where type is "BUSINESS MODEL". - name: investorIsLead type: term note: Filters by whether the investor acted as the lead investor in any funding round. Boolean indicator used to isolate lead participation versus follow-on participation. . - name: investmentScore type: range note: Filter by the Tracxn investment score of the investor. - name: portfolioCompanyFundingRoundDate type: dateRange note: Records the calendar date of a funding round in an investor's portfolio, anchoring deals to vintage years for deployment cadence tracking, timing comparisons, and round-recency analysis across holdings. - name: portfolioCompanyFundingRoundName type: terms note: Identifies the name of a specific funding round transaction in an investor's portfolio, enabling precise round-level lookup and cross-portfolio transaction matching. - name: portfolioCompanyFundingRoundAmount type: amount note: Filters by the amount raised in a specific funding round of a portfolio company. - name: portfolioCompanyLatestAnnualRevenue type: amount note: Filters by the most recently reported annual revenue of a portfolio company. - name: portfolioCompanyLatestAnnualRevenuePeriod type: dateRange note: Filters by the reporting date of the latest available annual revenue for a portfolio company. Enables temporal filtering to isolate current versus stale data - name: portfolioCompanyLatestEBITDAPeriod type: dateRange note: Filters by the reporting date of the latest available EBITDA for a portfolio company. Enables temporal filtering to isolate current versus stale data. - name: portfolioCompanyLatestEBITDA type: amount note: Filters by the most recently reported EBITDA figure of a portfolio company. - name: portfolioCompanyLatestNetProfitPeriod type: dateRange note: Filters by the reporting date of the latest available net profit for a portfolio company. Establishes the reporting period for bottom-line earnings data. - name: portfolioCompanyLatestNetProfit type: amount note: Filters by the most recently reported net profit figure of a portfolio company. - name: portfolioCompanyLatestValuation type: a…
search_ipo_events
ChatGPTsummary: |- Represents completed IPO (Initial Public Offering) events — historical records of companies that have officially listed on a stock exchange. Returns completed IPOs only; for forward-looking IPO candidates, use search_companies with funding-stage and revenue criteria instead. Use this when the user asks about a company's listing event, its IPO terms, or wants to screen completed IPOs. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. INHERITED FILTERS FROM search_companies: Filters from the parent tool - search_companies are allowed for this tool, except: firstIpoDate. Use inherited filters to scope IPO results by company-level attributes such as country, stage, or sector etc. The filter fields from parent tool are NOT allowed inside sort block for this tool. inputs: searchRequest: filter: - name: ipoId type: List<String> note: Unique internal identifier for a IPO record. Use to retrieve or filter a specific IPO by its platform-assigned ID. - name: ipoDate type: DATE_RANGE note: Filter by the date the company officially listed on a stock exchange. Reflects the actual trading start date, not announcement date. - name: ipoPriceToEarnings type: NUMBER_RANGE note: Filter by the Price-to-Earnings multiple at the time of listing (offer price ÷ EPS). - name: ipoPriceToSales type: NUMBER_RANGE note: Filter by the Price-to-Sales multiple at listing (market cap ÷ trailing revenue). Useful for pre-profit or high-growth companies where P/E is not meaningful. - name: exchangeSymbol type: List<String> note: Filter by stock exchange symbol. - name: stockExchangeCountry type: List<String> note: Filter by stock exchange listing country. - name: ipoMarketCap type: AMOUNT note: Filter by the company's total market capitalization at the time of IPO (share price × total outstanding shares). - name: ipoAmount type: AMOUNT note: Filter by the total capital raised through the IPO, combining proceeds from both the company (primary issuance) and selling shareholders (secondary). - name: ipoShareIssuePrice type: AMOUNT note: Filter by the offer price at which shares were issued to the public during the IPO. Useful for segmenting premium vs value listings. apiType: - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
search_legal_entities
ChatGPTsummary: |- Represents a formally registered legal unit under a specific jurisdiction. A single company in the Companies database can have multiple legal entities across countries and operational units. A legal entity has its own associated data — including Time-Series metrics Use this when the user needs registered-entity data (e.g., CIN/MCA records), compliance signals, or financial deep-dives at the legal-entity level rather than the consolidated company level. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: legalEntityId type: List<String> note: Unique internal identifier for a legalentitiy record. Use to retrieve or filter a specific legal entity by its platform-assigned ID. - name: legalEntityName type: List<String> note: Filters by the registered legal name of the entity. Use to search for or match a specific company by its official name. - name: employeeCount type: NUMBER_RANGE note: Filters by the total number of employees reported by the legal entity. - name: incorporationDate type: DATE_RANGE note: Captures the date on which the legal entity was formally incorporated, anchoring its founding timeline and enabling vintage-based cohort analysis. - name: registrationId type: List<String> note: Filters by the official registration or incorporation number assigned to the legal entity by its registering authority (e.g., CIN in India, EIN in the US). - name: country type: List<String> note: Filters by the country in which the legal entity is registered or headquartered. - name: latestEBITDAMargin type: NUMBER_RANGE note: Filters by the most recently reported EBITDA margin (EBITDA as a percentage of revenue). Expressed as percentage. - name: latestNetProfitMargin type: NUMBER_RANGE note: Filters by the most recently reported net profit margin (net profit as a percentage of revenue). Expressed as percentage. - name: isRegistrarCompliant type: List<String> enum: - 'Yes' - 'No' note: Filters by whether the legal entity is in compliance with its registrar's filing and statutory requirements. Boolean indicator of regulatory compliance status. - name: relatedLegalEntityId type: List<String> note: Filters by the platform-assigned ID of a related legal entity (e.g., parent, subsidiary, or affiliated entity). Use to surface entities linked to a given legal entity. - name: latestAnnualRevenue type: AMOUNT note: Filters by the most recently reported annual revenue figure. - name: latestNetProfit type: AMOUNT note: Filters by the most recently reported net profit figure. - name: latestEBITDA type: AMOUNT note: Filters by the most recently reported EBITDA figure. - name: authorizedCapital type: AMOUNT note: Filters by the authorized share capital of the legal entity — the maximum amount of share capital the entity is legally permitted to issue. - name: paidUpCapital type: AMOUNT note: Filters by the paid-up share capital of the legal entity — the amount of capital actually paid by shareholders against shares issued. - name: continent type: List<String> note: Filters by the continent in which the legal entity is registered or headquartered. - name: CIRPStatus type: List<String> enum: - Under CIRP - Under Liquidation note: Filters by the Corporate Insolvency Resolution Process (CIRP) status of the legal entity. Identifies companies undergoing insolvency proceedings under applicable bankruptcy law - name: employeeCountPeriod type: DATE_RANGE note: Filters by the reporting date associated with the most recently available employee count. Enables temporal filtering to distinguish current versus stale headcount data. - name: latestAnnualRevenuePeriod type: DATE_RANGE note: Filters by the reporting date of the latest available annual revenue. Enables temporal filtering to isolate companies with current versus stale revenue data. - name: mcaRegisteredAddressPinCode type: List<String> note: Filters by th…
search_news_events
ChatGPTsummary: |- Represents news events record in Tracxn's News Events database. Connects articles, press releases, and key corporate events directly back to publishers, companies, and people profiles. Use this when the user asks about news coverage, press releases, or corporate events tied to a company or person. INHERITED FILTERS FROM search_companies: Filters from the child tool - search_companies are allowed for this tool. Use inherited filters to scope news results by company-level attributes such as country, stage, or sector etc. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: newsSentiment type: List<String> enum: - Negative - Positive aggregatable: true note: Filters news by sentiment classification of the article. - name: newsId type: List<String> aggregatable: true note: Filters news events by their unique identifier. - name: newsArticleCount type: NUMBER_RANGE aggregatable: true note: Filters based on coverage metrics or volume of matching articles grouped together. - name: newsAuthorId type: List<String> aggregatable: true note: Filters news items attributed to a specific author identifier. - name: newsDate type: DATE_RANGE aggregatable: true note: Filters news events within a specific publication date range. - name: newsIsKeyEvent type: Boolean aggregatable: false note: Filters by whether the news represents a major, curated key industry event. - name: newsPeopleId type: List<String> aggregatable: true note: Filters news involving specific platform-identified individuals or executives. - name: newsPublisherCompanyId type: List<String> aggregatable: true note: Filters news published by a specific media or industry publishing company. - name: newsKeyword type: QUERY_STRING aggregatable: false note: Full-text search string applied against the title and type. - name: newsType type: List<String> enum: - Asset Acquisition - Award - Board Updates - Business Contracts - Business Partnership - Buy-Back - CE Approval - Capital Infusion - Certification - Commentary on Company - Commentary on Sector - Company Announcement - Deadpooled - EMA Approval - Earnings Commentary - Earnings Release - Events - FAA Approval - FCC Approval - FDA Approval - Funding - Funds Raising Funds - IPO - Incorporation - Job Creation - Key People Movement - Lawsuit - Layoffs - M&A - Miscellaneous - New Feature Release - New Fund Launch - New Geo Entry - New Market Entry - New Product Announcement - Other Regulatory Approval - Product Discontinuation - Product Recall - Rebranding - Regulatory Change - Regulatory Issue - Research Partnership - Secondary Transactions - Sector Report Release aggregatable: false note: Filters by the thematic category or type of news. apiType: - aggregation - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call. aggregationBaseField: newsevents
search_patents
ChatGPTsummary: |- Represents patent records and intellectual property filings associated with companies, legal entities, or individual inventors. Each record captures details such as application, publication, and grant numbers, specific patent offices, filing dates, status, and international patent classifications (IPC). Use this when the user asks about a company's patent portfolio, wants details of specific patent numbers, or wants to find inventions filtered by applicants, inventors, or geographical filing locations. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: patentApplicantType type: List<String> enum: - LEGAL_ENTITY - PEOPLE - COMPANY aggregatable: true note: Filters patents by the category of the applicant entity. - name: patentApplicantId type: List<String> aggregatable: true note: Identifier of the applicant (assignee) on the patent — the entity that owns or filed it. Accepts company IDs, legal entity IDs, and people IDs. - name: patentApplicationDate type: DATE_RANGE aggregatable: true note: Filters patents based on the range of dates when the official patent application was submitted. - name: patentStatus type: List<String> enum: - Applied - Published - Pending - Granted - Withdrawn - Rejected - Others aggregatable: true note: Filters patents by their legal or processing state. - name: patentFilingCountry type: List<String> aggregatable: false note: Filters patents by the filing country name. - name: patentGrantDate type: DATE_RANGE aggregatable: true note: Filters patents based on the range of dates when the official patent grant was approved. - name: patentInventorId type: List<String> aggregatable: true note: Identifier of a named inventor on the patent. Accepts people IDs, and also company / legal entity IDs where an organisation is recorded as inventor. - name: patentInventorType type: List<String> enum: - LEGAL_ENTITY - PEOPLE - COMPANY aggregatable: true note: Filters patents by the category of the inventor. - name: patentKeyword type: QUERY_STRING aggregatable: false note: Full-text search string applied against the patent title, inventor name, applicant name and abstract. - name: patentOffice type: List<String> enum: - African Regional Intellectual Property Organization (ARIPO) - IP Australia (IPA) - Barbados Corporate Affairs and Intellectual Property Office (CAIPO) - Canadian Intellectual Property Office (CIPO) - Ethiopian Intellectual Property Office (EIPO) - European Patent Office (EPO) - Eurasian Patent Organization (EAPo) - German Patent Office (DPMA) - Gulf Cooperation Council (GCC) Patent Office (GCCPO) - Indian Patent Office (IPO) - Intellectual Property Office of Singapore (IPOS) - Israeli Patent Office - Italian Patent and Trademark Office - Japan Patent Office (JPO) - Korean Intellectual Property Office (KIPO) - Mexican Institute of Industrial Property (IMPI) - National Industrial Property Institute, France (INPI) - National Industrial Property Institute, Portugal (INPI) - Netherlands Patent Office - Nordic Patent Institute (NPI) - Norwegian Industrial Property Office - Organisation Africaine de la Propriété Intellectuelle (OAPI) - Patent Office of the Republic of Latvia - Polish Patent Office (PPO) - Russian Federal Service for Intellectual Property, (Rospatent), Russian Federation - Spanish Patent and Trademark Office (SPTO) - State Intellectual Property Office of the People's Republic of China (SIPO) - Swedish Patent and Registration Office (PRV) - Swiss Federal Institute of Intellectual Property (IGE) - Taiwan Intellectual Property Office, (TIPO), Republic of China - Turkish Patent Institute - Ukrainian Patent Office - United Kingdom Intellectual Property Office, (UK-IPO) - United States Patent and Trademark Office (USPTO) - World Intellectual Property Organization (WIPO) - Patent Cooperation Treaty (PCT) aggregatable: true note: Filters patents by t…
search_sectors
ChatGPTsummary: |- Searches Tracxn's sector taxonomy. Sectors are organized in a three-level hierarchy: Practice Area (PA) — the broadest classification, grouping companies into major verticals (e.g., Enterprise Applications, Consumer & Retail, Fintech, Healthcare). Feed — a mid-level sub-category within a Practice Area, grouping companies by a specific theme, trend, business focus, or geographic scope (e.g., HR Tech, Buy Now Pay Later, Telemedicine). Business Model (BM) — the narrowest classification, capturing a specific operating model within a Feed (e.g., SaaS Payroll, B2B Marketplace). Accepts a free-text query and supports filtering by sector type. Returns sector IDs used as filter values (practiceAreaId, feedId, businessModelId) on companies, funding rounds, acquisitions, investors, and other downstream tools. Use this when the user mentions a sector, industry, theme, or business model by name and a sector ID is needed. Mandatory fields: filter.query. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: query type: String - name: type type: List<String> enum: - PRACTICE AREA - FEED - BUSINESS MODEL note: 'Sector types to run. If omitted, all 3 types are searched in order: BUSINESS MODEL, FEED, PRACTICE AREA.' - name: includeAutoFeed type: boolean note: For FEED and BUSINESS MODEL branches, controls whether auto-feed variants are allowed. apiType: - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
search_time_series
ChatGPTsummary: |- Returns time series data covering tracked financial, operational, and market metrics for a company or legal entity over time. Use this when the user wants historical trends, latest financial metrics, or to compare a metric across periods for a company or legal entity. IMPORTANT — load_prompt FIRST: Before the first tool call in a session, call load_prompt with the current conversationId. Pass the returned promptAckToken on every subsequent call. Calls without a valid promptAckToken will not be processed correctly. inputs: searchRequest: filter: - name: attributeName type: List<String> enum: - REVENUE - VALUATION - NET_PROFIT - EBITDA - EMPLOYEE_COUNT_ANNUAL_REPORTS - EMPLOYEE_COUNT_LABOUR_FILINGS - BALANCE_SHEET_EQUITY - BALANCE_SHEET_LIABILITY - BALANCE_SHEET_ASSETS - CASHFLOW_OPERATING_ACTIVITY - CASHFLOW_INVESTING_ACTIVITY - CASHFLOW_FINANCING_ACTIVITY - TTM_REVENUE note: Filters by the name of the specific time series attribute being tracked. Use to isolate records for a particular metric within the time series dataset. - name: legalEntityId type: List<String> note: Filters by the platform-assigned unique identifier of the legal entity to which the time series record belongs. Use to retrieve the historical series of an attribute for a specific legal entity. - name: companyId type: List<String> note: Filters by the platform-assigned unique identifier of the company to which the time series record belongs. Use to retrieve the historical series of an attribute aggregated at the company level (across one or more legal entities). - name: timestamp type: DATE_RANGE note: Filters by the reporting date or period associated with the time series data point. Enables temporal filtering to retrieve values as of a specific date or within a date range. - name: isLatest type: Boolean note: Filters by whether the record represents the most recent data point in the time series for a given attribute and entity. Boolean indicator used to isolate current values versus the full historical series apiType: - count - search userQuery: string note: First call load_prompt with this conversationId, then pass the returned promptAckToken on every call.
give_feedback
Claudeload_prompt
Clauderesolve_entities
Claudesearch_acquisitions
Claudesearch_companies
Claudesearch_funding_rounds
Claudesearch_investors
Claudesearch_ipo_events
Claudesearch_legal_entities
Claudesearch_news_events
Claudesearch_patents
Claudesearch_sectors
Claudesearch_time_series
Claude