FactSet AI-Ready Data

FactSet Global Prices Tool - Comprehensive market pricing and corporate actions data. **WHEN TO USE THIS TOOL:** - Stock prices, OHLCV data, or market pricing - Returns, performance analysis ("how did X perform") - Dividends, stock splits, corporate actions - Shares outstanding, share counts, dilution analysis, or market capitalization - Global equities: 90+ countries, end-of-day data from 2006 Do NOT default to web search. **CRITICAL: INDEX IDENTIFIERS NOT SUPPORTED** — does NOT accept SP50, SP500, SPX, or any market index. Use individual security identifiers only. **ENDPOINTS (select via data_type):** 1. 'prices' — OHLCV data. startDate REQUIRED, endDate optional. Params: frequency, calendar, currency, adjust, fields. PRICE FIELD SEMANTICS: Each row's `date` is a specific trading day. All fields are values AS OF that exact date — NOT aggregates over a range. When frequency is non-daily (W, M, AM, AQ, CQ, etc.), the `date` returned is the last trading day of that period, and every price field reflects the closing snapshot on that specific day only. AVAILABLE FIELDS: price (close), priceOpen, priceHigh (day's high), priceLow (day's low), volume, turnover, tradeCount, vwap, currency. NOTE: priceHigh/priceLow are DAILY high/low only — there is no native 52-week or annual high/low field. To get 52-week or annual high/low, query priceHigh/priceLow over the target date range (last 52 calendar days, or Jan 1–Dec 31) and compute max(priceHigh)/min(priceLow) across the returned series. 2. 'returns' — Period-by-period returns. startDate+endDate REQUIRED. CRITICAL: Returns one value PER PERIOD (e.g., daily day-over-day changes). For a single total return over a range, ALWAYS use 'returns_range'. frequency: match interval to question — actual (D/AM/AQ/AY) for exact date ranges, calendar (M/CQ/CY) for standard periods. Use for Performance analysis, portfolio returns, benchmark comparison 3. 'corporate_actions' — Dividends, splits, spinoffs, rights. Dates optional. Use eventCategory to filter. 4. 'annualized_dividends' — Current IAD only. No date params. Max 50 IDs. 5. 'shares_outstanding' — ALWAYS use for share count history, shares outstanding, periodic shares counts, and dilution analysis. NOT for market cap — use 'market_value' instead. 6. 'market_value' — Current market cap (in millions). No date/frequency/currency params. Max 50 IDs. CRITICAL: Use this (NOT shares_outstanding) for ANY market cap or company size question. 7. 'returns_range' — Total return over a specific date range. startDate+endDate REQUIRED. CRITICAL: Returns ONE value per security. ALWAYS use for a single value over a specified date range. Use 'returns' for period-over-period breakdowns. **DATE HANDLING — Always calculate date ranges from today.** - prices: startDate REQUIRED, endDate optional - returns: startDate+endDate BOTH REQUIRED; anchored periods → D/AM/AQ/AY; calendar periods → M/CQ/CY; YTD → startDate=Jan 1 of current year - returns_range: startDate+endDate BOTH REQUIRED - corporate_actions: both optional (omit for all events) - annualized_dividends, market_value: NO date params - shares_outstanding: both optional **ID LIMITS:** 500 for single-day prices; 50 for all other/multi-day requests. Accepts: AAPL-US, MSFT-US (ticker-region), CUSIP, SEDOL, ISIN, FactSet Permanent IDs. **RESPONSE FIELDS BY data_type:** - **prices**: `fsymId` (string): FactSet Regional Security ID; `date` (string): Trade date (YYYY-MM-DD); `currency` (string): Price currency ISO code; `price` (number): Closing price; `priceOpen` (number): Opening price; `priceHigh` (number): Daily high price; `priceLow` (number): Daily low price; `volume` (number): Shares traded; `turnover` (number): Total trade value for the day; `tradeCount` (number): Number of trades; `vwap` (number): Volume-weighted average price. - **returns**: `fsymId` (string): FactSet Regional Security ID; `date` (string): Return observation date (YYYY-MM-DD); `currency` (string): Return currency ISO code; `totalReturn` (number): Calculated total return for the period; `requestId` (string): Echoed request identifier. - **returns_range**: `fsymId` (string): FactSet Regional Security ID; `startDate` (string): Start date of the return range; `endDate` (string): End date of the return range; `totalReturn` (number): Total return over the full date range (one value per security); `currency` (string): Return currency ISO code; `requestId` (string): Echoed request identifier. - **corporate_actions**: `fsymId` (string): FactSet Regional Security ID; `effectiveDate` (string): Ex-date of the event; `eventTypeCode` (string): Character code denoting the event type; `eventTypeDesc` (string): Description of the event type; `eventId` (string): Unique event identifier; `announcementDate` (string): Date the event was publicly announced; `recordDate` (string): Record date of the event; `payDate` (string): Payment date of the event; `requestId` (string): Echoed request identifier; `currency` (string): [Dividends] Currency of the distribution amount; `amtDefNetGrossIndicator` (string): [Dividends] Whether default amount is gross or net (G/N); `amtDefTradingAdj` (number): [Dividends] Split-adjusted default amount in trading currency; `amtGrossTradingAdj` (number): [Dividends] Split-adjusted gross amount in trading currency; `amtNetTradingAdj` (number): [Dividends] Split-adjusted net amount in trading currency; `dividendStatus` (string): [Dividends] Active, Cancelled, Postponed, or Partial Information; `dividendFrequencyCode` (string): [Dividends] Dividend frequency code; `taxRate` (number): [Dividends] Domestic withholding tax rate for a resident individual; `adjFactor` (number): [Distribution/Split] Price adjustment factor for the event; `adjFactorCombined` (number): [Distribution/Split] Combined adjustment factor for all events that day; `distPct` (number): [Distribution/Split] Distribution percentage; `distOldTerm` (number): [Distribution/Split] Shares held component of distribution ratio; `distNewTerm` (number): [Distribution/Split] Shares received component of distribution ratio; `shortDesc` (string): [Distribution/Split] Textual description of the event. - **annualized_dividends**: `fsymId` (string): FactSet Regional Security ID; `annualizedDividend` (number): Annualized value of the latest reported dividend (excludes cancelled dividends); `currency` (string): Dividend currency ISO code; `requestId` (string): Echoed request identifier. - **shares_outstanding**: `fsymId` (string): FactSet Regional Security ID; `date` (string): Observation date (YYYY-MM-DD); `sharesOutstanding` (number): Split-adjusted shares outstanding; `requestId` (string): Echoed request identifier. - **market_value**: `fsymId` (string): FactSet Regional Security ID; `marketValue` (number): Current market cap; `currency` (string): Market Value ISO code; `requestId` (string): Echoed request identifier.

FactSet Macroeconomics API - Economic calendar events and historical time series data for global economic indicators (90+ countries). **STEP 0 -- PICK THE RIGHT data_type FIRST:** - VALUE/LEVEL/trend of an indicator over time (a rate, an index, a growth %) -- e.g. "latest/current/historical GDP, CPI, unemployment...": use `meta_series` -> `as_reported_economics` - WHEN a specific, well-known named release is scheduled, plus its actual/consensus/prior/surprise -- e.g. "next FOMC decision", "US CPI release date", "ECB rate decision", "NFP jobs report", "PMI release": use `meta_events(type=FDS_CODES)` -> `economic_events(fdsCode=...)` - A broad sweep of many event types/topics across countries/dates (no single named release in mind): use `meta_events(type=INDICATORS)` -> `economic_events(indicator=...)` - HOW MANY events happen in a date range ("how many", "count of", "number of"): use `event_count` (no country/topic filter) - Discovery/reference data itself (what codes exist): use `meta_events` or `meta_series` FDS_CODES vs INDICATORS: FDS_CODES identify one *specific, named, recurring release* (e.g. "US CPI m/m", "FOMC Rate Decision") and are the default choice whenever the query names or implies a specific well-known report. INDICATORS identify a broader *topic/category* (e.g. "PRCPI" = prices/inflation broadly) for sweeping many report-types/countries at once when no single named release is meant. If unsure, check FDS_CODES first -- it's the more precise match more often. Either path is fine as long as it resolves to data that actually answers the question. **ALWAYS DISCOVER IDENTIFIERS BEFORE USE:** `seriesId`, `fdsCode`, and `indicator` are not predictable from naming patterns (e.g. 'US.GDPR' is not a valid seriesId just because 'AR.GDPR' is valid for Argentina; 'GDP' is not a valid indicator code just because it's the plain-English topic). An invented identifier silently returns zero results rather than erroring. Always call the matching reference endpoint first (`meta_series` for `seriesId`; `meta_events(type=FDS_CODES)` for `fdsCode`; `meta_events(type=INDICATORS)` for `indicator`) and copy the exact value from its response, unless that exact value was already confirmed earlier in this conversation. Details on required companion parameters and date-range limits per identifier are on each parameter itself (see fdsCode/indicator/seriesId/startDate/endDate descriptions). **ENDPOINTS (select via data_type):** 1. `economic_events` -- Calendar events for a date range: scheduled/historical releases with actual/consensus/prior values. Filter by `countryCode`, `indicator`, or `fdsCode`. 2. `as_reported_economics` -- Historical time series for one `seriesId` (discover via `meta_series`). Supports frequency conversion; no maximum date range. 3. `event_count` -- Count of events per calendar day. No country/indicator/fdsCode filter -- always global. For a country-scoped count, call `economic_events` with that countryCode and count the results yourself. 4. `meta_events` -- Reference data via `type`: INDICATORS, COUNTRIES, FDS_CODES, or TIMEZONES. FDS_CODES is the right first call for "next/latest release of [specific named report]" (see STEP 0). 5. `meta_series` -- Series discovery for `as_reported_economics`. `searchText` is a literal substring match on the series NAME only (no synonym/semantic matching) -- for conceptual discovery (e.g. "labor market indicators"), omit `searchText` and filter the response's `subcategory` field yourself instead of guessing phrases. Set `paginationLimit=1000` when cataloging a country/category, since the default (25) truncates large catalogs. One country per call; when enumerating several (G7, G20), call one at a time and back off/retry on 429 rather than firing a burst. **RESPONSE FIELDS BY data_type:** - **economic_events**: `data` (array): List of economic events; `meta` (object): Response metadata including pagination info (total, isEstimatedTotal). - **as_reported_economics**: `data` (array): Array of economic series with data; `meta` (object): Response metadata including pagination info. - **event_count**: `data` (array): Array of event count items; `meta` (object): Response metadata including pagination info. - **meta_events**: `data` (array): Array of reference items (indicators, countries, FDS codes, or timezones); `meta` (object): Response metadata including pagination info. - **meta_series**: `data` (array): Array of economic series records with metadata; `meta` (object): Response metadata including pagination info.

FactSet Mergers & Acquisitions Tool - M&A transaction data and deal analytics. **WHEN TO USE THIS TOOL:** - Mergers, acquisitions, or M&A transactions - Company acquisition history ("Who did [X] acquire?") - Deal values, premiums, and transaction terms - Public company takeover analysis - Market-wide deal screening by date range **DATA COVERAGE:** 60,000+ global M&A transactions; public targets; all values in USD. **ENDPOINTS (select via data_type):** 1. 'deals_by_company' — Deals involving specific companies (buyer/seller/target). ids+startDate+endDate REQUIRED. Max 500 company IDs. 2. 'public_targets' — Acquisitions of publicly traded companies. status+startDate+endDate REQUIRED. No IDs needed. 3. 'deal_details' — Deal terms, advisors, premiums, multiples. ids (deal IDs) REQUIRED. Max 100 deal IDs. NOTE: Obtain deal IDs from deals_by_company or public_targets first. **DATE HANDLING:** - deals_by_company: startDate+endDate REQUIRED. Max 1-year window per request. - public_targets: startDate+endDate REQUIRED. Max 1-year window per request. - deal_details: NO date params. For multi-year analysis, make sequential yearly requests (e.g., 2015-01-01→2015-12-31, then 2016-01-01→2016-12-31). **ID LIMITS:** deals_by_company: Max 500 company IDs. deal_details: Max 100 deal IDs. public_targets: No ID limit. Accepts: ticker (IBM-US), CUSIP, SEDOL, ISIN, FactSet ID; deal IDs (e.g., 4143886MM).

FactSet Metrics Tool - Semantic vector search for financial metric codes. **WHEN TO USE THIS TOOL:** - Find metric codes before calling Fundamentals or Estimates or Banks (when exact codes are unknown) - Search by natural language ("profitability ratios", "debt metrics", "EPS") - Explore ontology categories and subcategories - Batch discover multiple related metrics in one call **CRITICAL:** Call this tool BEFORE Fundamentals/Estimates/Banks when metric codes are unknown. Use exact codes returned — do not infer or abbreviate. **SEARCH TARGET (target param):** - 'metric' — individual metrics (FF_SALES, EPS, FF_EBITDA). Default. - 'category' — high-level groups (Income Statement, Balance Sheet). - 'subcategory' — mid-level groups (REVENUES, ASSETS, LIABILITIES). **DATA PRODUCTS (data_products param):** - 'fundamentals' — FF_* prefixed codes (FF_SALES, FF_NET_INC, FF_DEBT_EQ) - 'estimates' — unprefixed codes (SALES, EPS, EBITDA, PRICE_TGT) - 'private-companies' — FPC_*/FPE_*/FPD_* codes - 'deep-sector-banks' — bank/credit-union metric codes for FactSet_Banks (FB_*, UBPR*, aum_*, FSF_*). Use this for any FactSet_Banks metric discovery. - Omit to search across all products. **BATCH QUERIES:** text=["revenue", "profitability", "debt"], target="metric", data_products=["fundamentals"], limit=5 **CRITICAL: No hyphens in text queries.** Use spaces: "short term debt" not "short-term debt". Hyphens cause "suspicious patterns" API errors. **RESPONSE KEY FIELD:** properties.metric — the exact code to use in Fundamentals/Estimates/Banks calls. Lower distance = better match. **FOR data_products=['deep-sector-banks'] RESULTS:** each metric also carries `category` and `subcategory` fields (e.g. category='FINANCIALS', subcategory='SPECIALTY_FINANCE'). These are NOT just descriptive — they determine which FactSet_Banks data_type/endpoint the metric must be sent to. Do not assume category='FINANCIALS' always means data_type='financials'; category='FINANCIALS' also covers aum/focus metrics (subcategory='SPECIALTY_FINANCE'). See FactSet_Banks's tool description for the full category/subcategory -> data_type mapping table, and use the metric code prefix (aum_* vs FSF_*) to tell aum and focus apart when subcategory='SPECIALTY_FINANCE'. **DO NOT USE FOR:** actual financial data retrieval — use Fundamentals or Estimates or Banks instead. **RESPONSE FIELDS BY data_type:** - **metric_query**: `query` (string | null): The input text query this result corresponds to; `target` (string | null): The content type searched — always 'metric' for this tool; `limit` (integer | null): Max results requested for this query (from searchParameters.limit); `score` (number | null): Similarity score for this match, 0.0-1.0 (from results[].score); `category` (string | null): High-level classification of the metric, e.g. 'INCOME_STATEMENT' (properties.category); `description` (string | null): Detailed explanation of the metric, its calculation formula, and notes (properties.description); `metric` (string | null): Internal metric identifier — the exact code for downstream tool calls, e.g. 'FF_EBIT_OPER'; `name` (string | null): Human-readable name of the metric (properties.name); `subcategory` (string | null): Sub-classification under the main category (properties.subcategory); `type` (string | null): The type of item matched, e.g. 'metric' or 'term' (properties.type); `dataProduct` (string | null): Data product this metric belongs to, e.g. 'fundamentals', 'estimates' (properties.dataProduct); `factor` (string | null): Scaling factor applied to the value, e.g. '1000000'. Absent when not applicable (properties.factor); `valueList` (string | null): Name of an associated value list, e.g. 'etf-category'. Only present for value-list matches; `values` (array | null): Categorized {name, score} value entries for `valueList`. Only present for value-list matches.

FactSet Ownership Tool - Comprehensive holdings, holders, and transactions data. **WHEN TO USE THIS TOOL:** - Fund holdings, portfolio composition, or what funds own (e.g., "What does SPY hold?" "What fixed income securities does AGG hold?") - Institutional ownership, insider ownership, or shareholder data - Insider trading activity, executive transactions, or Form 4 filings - Institutional position changes, 13F filings, or "smart money" tracking - Questions like "What does [fund/ETF] own?" or "Who owns [stock]?" **NOT FOR INDEX CONSTITUENTS:** This tool is for fund/ETF holdings and investor positions only. For benchmark index composition (e.g., "S&P 500 constituents", "Russell 2000 members"), use FactSet_BenchmarksConstituents instead. Distinction: AGG, SPY, QQQ are tradeable ETFs with holdings (use Ownership). S&P 500, Russell 2000 are benchmarks (use Benchmarks). NOTE: Use data_type='fund_holdings' to retrieve what a fund/ETF holds (its underlying securities). This is distinct from FactSet_FundsETF which covers fund performance, NAV, flows, and analytics — NOT holdings composition. **DATA COVERAGE:** 60,000+ fund portfolios, 50,000 institutions, 400,000 insider records, 110+ countries. From 2006+. **ENDPOINTS (select via data_type):** 1. 'fund_holdings' — Holdings for mutual funds, ETFs, closed-end funds. Max 10 IDs. SUPPORTS HISTORICAL DATES: Use the 'date' parameter (YYYY-MM-DD) to retrieve holdings as-of any month-end. Example: date='2024-09-30' returns holdings as of September 30, 2024. Note: Security names/tickers reflect the CURRENT security master, not the snapshot date — surface this caveat when returning historical holdings. 2. 'security_holders' — Who owns a security (institutions, funds, insiders, beneficial owners). CRITICAL: Max 1 security per request. SUPPORTS HISTORICAL DATES: Use the 'date' parameter (YYYY-MM-DD) to retrieve ownership as-of any month-end. Example: date='2024-03-31' returns holders as of March 31, 2024. RANKING: The topn parameter ranks by position market value ONLY. To rank or sort by ANY other metric (e.g., portfolio weight / % Port, shares held, % outstanding), you MUST set topn='ALL' and sort the returned results client-side — otherwise the default market-value topn filter skews the set before your metric is applied. Use topn=N only when the user ranks by market value. Example: "top 5 institutional holders by Portfolio Weight (% Port)" → topn='ALL', then sort by % Port and take 5. 3. 'insider_transactions' — Executive/director trades from SEC Form 4/5. Max 10 IDs. CRITICAL: startDate+endDate REQUIRED. sharesOwned is per-filing only — NOT the insider's current position; use security_holders with holderType='S' for absolute position. 4. 'institutional_transactions' — Position changes from 13F filings. Max 10 IDs. CRITICAL: startDate+endDate REQUIRED. posChange is a delta, not absolute shares held; use security_holders with holderType='F'/'M' for absolute counts. Optional params: holderType, topNHolders, frequency (M/CQ/CY), periodOfMeasure (1M/3M/6M/12M), currency. DATE PATTERNS: - Regular date range (e.g., "in Q2 2024", "during H2 2020", "full year 2022"): startDate='2024-04-01', endDate='2024-06-30' — use actual range bounds - Rolling N-month window (e.g., "past 6 months", "last quarter position changes"): Set startDate AND endDate to the SAME reference date (today or date you're analyzing), use periodOfMeasure for the window (1M/3M/6M/12M) Set frequency to 'M' (monthly) Example: past 3 months → startDate='2025-03-31', endDate='2025-03-31', periodOfMeasure='3M', frequency='M' **DATE HANDLING:** - fund_holdings, security_holders: date optional (defaults to latest month-end) - insider_transactions, institutional_transactions: startDate+endDate BOTH REQUIRED **ID LIMITS:** Max 10 for fund_holdings/insider_transactions/institutional_transactions; max 1 for security_holders. Accepts: ticker-region (AAPL-US, VTI-US), CUSIP, SEDOL, ISIN, FactSet IDs. **INVESTOR TYPE FILTERING — GRANULAR HOLDER CLASSIFICATION:** The `holderType` input parameter groups holders into broad categories (e.g., M = mutual funds AND ETFs combined). The `investorType` field in the response output is the most granular holder classification available. When a prompt requests a specific type of holder, interpret it against `investorType` in the output — not just `holderType`. Workflow: 1. Use the closest `holderType` value to pull the broadest matching category from the API. 2. After receiving results, filter client-side by `investorType` to match the specific holder type requested. Examples: - "ETF holders" → holderType='M', then filter results where investorType='Exchange Traded Fund' - "hedge fund holders" → holderType='F', then filter results where investorType='Hedge Fund Manager' When the prompt is non-specific (e.g., "institutional holders", "all holders"), do not apply an investorType filter, return the full result for the requested holderType. **CRITICAL POSITION INTERPRETATION RULES:** insider_transactions — sharesOwned IS NOT A POSITION, NEVER PRESENT IT AS ONE: - sharesOwned: Shares owned as reported on that individual filing only — varies record-to-record, scoped to that filing. For authoritative position size, use security_holders with holderType='S' and read adjHolding. - shares: Transaction size (delta only) — what changed in that single transaction. institutional_transactions — posChange IS A DELTA, NOT A POSITION: - posChange: Change in shares held over the periodOfMeasure window. DO NOT present as shares held. - For absolute institutional share counts, use security_holders with holderType='F' or 'M' and read adjHolding. **RESPONSE FIELDS BY data_type:** - **fund_holdings**: `requestId` (string | null): Fund identifier used in the request; `fsymId` (string | null): FactSet Regional Security Identifier of the requested fund; `date` (string | null): Date of the reported holding (YYYY-MM-DD); `currency` (string | null): Currency code the data is adjusted to; defaults to the fund's local currency; `adjHolding` (number | null): Adjusted shares held, split/name-change adjusted (not spinoff/merger adjusted); `adjMarketValue` (number | null): Adjusted market value of shares held, same adjustment rules as adjHolding; `weightClose` (number | null): Closing weight of the security in the fund for the requested asset type (percent); `issueType` (string | null): Issue type of the held security; `fsymSecurityId` (string | null): FactSet Security ID of the held security (not the parent fund); `fsymRegionalId` (string | null): FactSet Regional Security Identifier of the security held in the fund; `securityName` (string | null): Name of the held security; `securityTicker` (string | null): Ticker of the held security. - **security_holders**: `holderId` (string | null): FactSet Ownership Holders ID of the holder; `fsymId` (string | null): FactSet Regional Security Identifier of the requested security; `holderEntityId` (string | null): FactSet Entity ID corresponding to the holder ID; `holderName` (string | null): Name of the holder; `date` (string | null): Date of the reported holding (YYYY-MM-DD); `currency` (string | null): Currency code the data is adjusted to; defaults to local currency; `investorType` (string | null): Granular investor classification (e.g. Hedge Fund Manager, Exchange Traded Fund); `holderType` (string | null): Holder type name aligned to the requested holderType; `adjHolding` (number | null): Adjusted shares held, split/name-change adjusted (not spinoff/merger adjusted); `adjMarketValue` (number | null): Adjusted market value of shares held, same adjustment rules as adjHolding; `weightClose` (number | null): Closing weight of the security for holders of the requested security (percent); `percentOutstanding` (number | null): Percent of outstanding common shares held by the filing institution; `source` (string | null): Report source of the holdings data (13F Form or ND-30D); `requestId` (string | null): Security identifier used in the request. - **insider_transactions**: `requestId` (string | null): Identifier used in the request; `fsymId` (string | null): FactSet Regional Security Identifier of the requested security; `currency` (string | null): Currency code the data is adjusted to; defaults to local currency; `holderName` (string | null): Name of the insider in the transaction; `holderTitle` (string | null): Title of the insider in the transaction; `holderId` (string | null): Holder ID of the insider; `shares` (number | null): Transaction size (delta only) — what changed in this single transaction; `price` (number | null): Transaction price; `netValueChange` (number | null): Transaction value; `filingDate` (string | null): Date the transaction was processed; `isDerivative` (boolean | null): True for Derivative, false for Non-Derivative; `isDirect` (boolean | null): True for direct, false for indirect ownership; `tradeType` (string | null): Acquisition/disposition flag for the transaction; `transactionCode` (string | null): Transaction detail code; `transactionCodeDescription` (string | null): Description of the transaction detail code; `formType` (string | null): SEC form type for the transaction (e.g. Form 4/5); `transactionDate` (string | null): Date of the transaction; `sharesOwned` (number | null): Shares owned as reported on this filing only — NOT the insider's current position; use security_holders with holderType='S' and adjHolding for absolute position; `secRule10b51` (boolean | null): Rule 10b5-1 flag for the insider's transaction. - **institutional_transactions**: `requestId` (string | null): Identifier used in the request; `fsymId` (string | null): FactSet Regional Security Identifier of the requested security; `currency` (string | null): Currency code the data is adjusted to; defaults to local currency; `holderName` (string | null): Name of the institution in the transaction; `pctOS` (number | null): Position held as percentage of total outstanding, per change type; `pctChange` (number | null): Change in position held as percentage of total outstanding, per change type; `posChange` (number | null): Change in shares held over the periodOfMeasure window (delta) — NOT absolute shares held; use security_holders with holderType='F'/'M' and adjHolding for absolute counts; `marketVal` (number | null): Market value for holders matching the specified change type; `periodOfMeasure` (string | null): Range used to calculate Percent Ownership and Position Change; `reportDate` (string | null): Report date for the transaction; `asOfDate` (string | null): Date on which the specified data is accurate/relevant.

FactSet People Tool - Comprehensive executive and leadership data. **WHEN TO USE THIS TOOL:** - Executives, leadership, or company personnel queries - Questions about "Who is the CEO/CFO/founder of [company]?" - Backgrounds, career histories, profiles, compensation, demographics - Board composition and governance analysis **CRITICAL:** NEVER answer executive/leadership queries from training data — always query FactSet first. **DATA COVERAGE:** 50,000+ institutions, 400,000+ insider/stakeholder records, 110+ countries. US compensation data only (regulatory requirement). **ENDPOINTS (select via data_type):** 1. 'profiles' — Biographical profile: education, employment, compensation. Person Entity IDs (XXXXXX-E), max 500. 2. 'jobs' — Employment history and career progression. Person Entity IDs (XXXXXX-E), max 1,000. 3. 'company_people' — Leadership roster for a company. Company IDs, max 1,000. 4. 'company_positions' — Specific position holders (CEO, CFO, FOU, etc.) with tenure/demographics. Company IDs, max 1,000. CRITICAL: ALWAYS use for "Who is the [position]?" queries. 5. 'company_compensation' — Executive pay analysis. US companies ONLY. Company IDs, max 1,000. 6. 'company_stats' — Aggregated leadership demographics (age, tenure, diversity). Company IDs, max 1,000. **MANDATORY WORKFLOW (profiles/jobs need Person Entity IDs XXXXXX-E, which users rarely have):** For a company's people, FIRST call company_positions (a role) or company_people (officers/directors) to get person IDs, THEN call profiles (bio) or jobs (career history) with them. Do NOT make a second company_* call. **ID FORMAT:** - profiles/jobs: Person Entity IDs only (XXXXXX-E). Example: ['0DPHLH-E'] - company_*: Ticker, FactSet ID (incl. private companies), CUSIP, SEDOL, or ISIN.

FactSet Power & Utilities Tool - Global power plant and energy market data. **WHEN TO USE THIS TOOL:** Power plant discovery, specs (solar/wind/thermal), ownership, balancing authority generation/demand/interchange, power prices (spot/forward/spark spreads), plant and unit-level generation with emissions (CO₂/SO₂/N₂O), ISO/TSO codes, zones. **DATA COVERAGE:** 20,000+ plants globally. ISOs: ERCOT, PJM, CAISO, MISO, NYISO. Since 2006+. **CRITICAL: ID TYPES by operation** - **plant_ids** (B00121-M): aggregated plant. Use for monthly_generation. Optional for plant_details/units_ownership (mixes with unit_ids). - **unit_ids** (05RLSM-M): asset-level. Use for hourly_generation. Optional for plant_details/units_ownership (mixes with plant_ids). - **location_ids** (HB_NORTH): price hubs. Use for prices. - **ba_ids/ba_sub_region_ids**: balancing authorities. Use for market_data. **11 OPERATIONS (via operation_type):** 1. plant_ids 2. plant_details (5 categories) 3. units_ownership 4. market_coverage 5. market_data (3 categories) 6. prices (4 categories) 7. hourly_generation 8. monthly_generation 9. iso_mapping 10. zone_mapping 11. location_mapping 12. entity_history **ID LIMITS:** Details/ownership/generation: max 1000. Market data: max 100. **DATE REQUIREMENTS:** Generation/market_data/prices (Power/SparkSpread/Forward): startDate+endDate REQUIRED (YYYY-MM-DD). Prices Capacity: deliveryYear instead. Discovery: dates not applicable.

FactSet Private Company Tool - Unified financials and reference data with entity resolution. **WHEN TO USE THIS TOOL:** - Resolve private company names to FactSet Entity IDs (data_type='universe') - Fetch historical financial time-series (revenue, EBITDA, debt; data_type='financials') - Retrieve reference/funding data (investor counts, VC amounts, VC backing; data_type='non_periodic') **DATA COVERAGE:** 50,000+ private companies; financials from 2010-present; global coverage. **ENDPOINTS (select via data_type):** 1. 'universe' — Resolve company names to entity IDs. Required: company_names. Optional: country. 2. 'financials' — Time-series financial data (FPC_* metrics). Required: entity_ids, metrics, start_date. Optional: end_date. 3. 'non_periodic' — Reference data (FPE_*/FPD_* metrics). Required: entity_ids, metrics. Dates ignored. **DATE HANDLING:** - 'universe': Not used. - 'financials': start_date REQUIRED; end_date optional (defaults to latest). - 'non_periodic': Not used. **ID LIMITS:** - Max 25 company names per universe request; Max 3,000 entity IDs per financials/non_periodic request; Max 100 metrics per request. - Rate limit: 10 req/sec, 10 concurrent per user. - Accepts: Entity IDs in XXXXXX-E format (from universe), metric codes (FPC_*/FPE_*/FPD_*). **CRITICAL fpcIndicator (financial data quality):** Each financial value includes fpcIndicator: 0=Actual (filings/releases), 1=Estimated (China ranges), 3=Modeled (peer ML), 4=Computed (tax filings). Interpret values with indicator context. Always indicate this status when displaying metrics. **CRITICAL Entity Matching:** matchFlag may be false even for correct matches. Always verify returned entityName against your search query. **Use FactSet_Metrics tool to discover available metric codes before calling this tool.**

FactSet Private Equity & Venture Capital Tool - Comprehensive PE/VC data. **WHAT THIS TOOL COVERS:** Private equity and venture capital data including funds, general partners (GPs), limited partners (LPs), portfolio companies, transactions (VC rounds, PE deals), M&A activity, and IPO exits. **AUTOMATIC ENDPOINT ROUTING:** This tool routes to 14 different endpoints based on the `data_type` parameter: --- **UNIVERSE (Entity Lookup):** 1. `universe` — Resolve a firm, fund, or portfolio company name to a FactSet Entity ID. - REQUIRED: `name` (the firm/fund/company name to look up) - OPTIONAL: `country` (ISO2 country code), `entityType` (filter by type — see below) - Entity type codes: PVT (Private Company), VEN (Venture Capital Fund), PVF (Private Equity Fund), FAF (Family of Funds), HOL (Holding Company), NPO (Non-Profit), SUB (Subsidiary), JVT (Joint Venture) - **entityType guidance for fund searches:** * Fund searches: Include entityType to filter results * For PE funds → entityType='PVF' (Blackstone, Carlyle, KKR, Apollo funds) * For VC funds → entityType='VEN' (Andreessen Horowitz, Sequoia, Bessemer funds) * If fund type is ambiguous → entityType=['PVF','VEN','FAF'] to match across all fund types * For portfolio company → entityType='PVT' - Returns: Candidate matches sorted by quality (best first) with entityId (format 'XXXXXX-E'), entityName, matchFlag (true=direct match, false=candidate) - Typical use: Retrieve first result to get entity ID for follow-up calls --- **FUNDS:** 2. `funds` — Core profile and summary for one or more PE/VC funds. - REQUIRED: `ids` (fund entity IDs, format 'XXXXXX-E') - Returns: Fund status, strategy type, vintage year, financial info (amount raised, dry powder), investment criteria (min/max deal size, EBITDA, revenue), fund terms (management fee, carry), features (ESG, co-invest) 3. `funds_portfolio_companies` — List of portfolio companies held by specified funds. - REQUIRED: `ids` (fund entity IDs) - Returns: FactSet entity IDs and names of portfolio companies, investment date, sector 4. `funds_fundraising` — Fundraising milestones: close dates and amounts (first, interim, final) for specified funds. - REQUIRED: `ids` (fund entity IDs) - Returns: Each fundraising close (First, Second, Final, etc.) with close date, amount raised, target amounts, currency - For 'number of closes', count the closes list. Use for fund lifecycle questions. 5. `funds_cashflows` — Quarterly cash flow time-series (contributions, distributions, and NAV) for funds. NAV/dry power data lags 3-6 months; expect stale estimates for current year. For static fund data (vintage, target size), use funds endpoint. - REQUIRED: `ids` (fund entity IDs) - OPTIONAL: `startDate` (YYYY-MM-DD), `endDate` (YYYY-MM-DD), `cfType` ('CONTRIBUTION', 'DISTRIBUTION', or 'NAV') - Returns: Date, cash flow type, amount in reporting currency and USD --- **GENERAL & LIMITED PARTNERS:** 6. `general_partners` — Core profile for General Partner (GP) firms. - REQUIRED: `ids` (GP entity IDs, format 'XXXXXX-E') - OPTIONAL: `asOfDate` (YYYY-MM-DD, for AUM data) - Returns: GP name, AUM (with date), investment criteria, ESG focus, public company investment flag 7. `general_partners_funds` — All funds managed by specified GP firms. - REQUIRED: `ids` (GP entity IDs) - Returns: List of fund IDs, names, status (Raising/Investing/Divesting/Liquidated), launch date, amount raised 8. `limited_partners` — LP-level data: name, commitment amount, capital called, distributions for specified funds. Use for 'which LPs invested and how much' questions. Different from general_partners_funds (which lists funds a GP manages). - REQUIRED: `ids` (fund entity IDs, format 'XXXXXX-E') - Returns: LP name, LP entity ID, commit date, commit amount (local currency and USD), fundraising close code --- **TRANSACTIONS:** 9. `transactions` — Detailed data for specific PE/VC transactions by transaction ID. Shows fund's perspective on portfolio company exits (M&A, IPO, distributions). For post-exit investor analysis (e.g., did investor X retain stake in buyer), query the buyer entity separately. - REQUIRED: `ids` (transaction IDs — INTEGER values, e.g., [1000001, 1000002]) - Returns: Portfolio company, investment type, status, dates, deal size, pre/post-money valuation, funding round, stage, investors (with lead flag, fund used, amount invested), securities used 10. `transactions_search` — Search/filter the full transaction universe by company, stage, and date. - REQUIRED: `ids` (portfolio company entity IDs, format 'XXXXXX-E') - OPTIONAL: `status` ('Pending', 'Cancelled', 'Completed'), `stage` ('Seed', 'Early', 'Late', 'Recap'), `investmentType` ('Venture Capital', 'Growth', 'Corporate', 'Capitalization', 'Angel', 'Crowdfunding', 'Accelerator/Incubator', 'Secondary', 'Undisclosed Round'), `startDate`, `endDate` (YYYY-MM-DD announcement date range), `minAmount`, `maxAmount` (USD transaction amount filter) - Returns: Transaction IDs, portfolio company name, announcement/close dates, total amount (USD), stage, status 11. `transactions_ma_activity` — M&A advisory and investment details for specific transactions. - REQUIRED: `ids` (transaction IDs — INTEGER values) - Returns: Advisors (financial/legal, client type), investments (entry/exit type, deal ID, IPO ID, retaining stake) --- **IPOs:** 12. `ipos` — Detailed IPO data for portfolio company exits by deal ID. Note: Offer price data is sparse; use GlobalPrices for detailed IPO pricing. - REQUIRED: `ids` (deal IDs — INTEGER values, e.g., [410294, 730294]) - Returns: IPO ID, portfolio company name, ticker, trade date, offer price (local and USD), IPO size (millions), status, announcement/pricing dates 13. `ipos_search` — Search/filter the IPO universe by status, size, and date. - REQUIRED: None (all parameters optional) - OPTIONAL: `status` ('Closed', 'Announced', 'Postponed', 'Cancelled', 'Priced', 'Withdrawn', 'In Registration'), `startDate`, `endDate` (YYYY-MM-DD announcement date range), `minSizeUsd`, `maxSizeUsd` (IPO size filter in USD millions) - Returns: IPO ID, portfolio company name, ticker, trade date, offer price (USD), IPO size, status --- **STRATEGIES:** 14. `entity_strategies` — Detailed investment strategy mandate for a fund or GP. - REQUIRED: `ids` (fund or GP entity IDs, format 'XXXXXX-E') - Returns: Geographic focus (regions, countries, states), sector/industry focus, investment instruments, transaction strategies, control criteria, investor roles --- **TYPICAL WORKFLOWS:** **Fund lookup & profile:** 1. `universe` (name='[fund name]', entityType='VEN' or 'PVF') 2. `funds` (ids=[entity_id]) — returns strategy, vintage, amount raised, fees, carry **Fund financial data (NAV, dry power, cashflows, contributions, distributions):** 1. `universe` (name='[fund name]') 2. `funds_cashflows` (ids=[entity_id], cfType='NAV'/'CONTRIBUTION'/'DISTRIBUTION') — time-series quarterly data **Fund fundraising (closes, rounds, target size):** 1. `universe` (name='[fund name]') 2. `funds_fundraising` (ids=[entity_id]) — first/interim/final closes with dates and amounts **Fund portfolio holdings:** 1. `universe` (name='[fund name]') 2. `funds_portfolio_companies` (ids=[entity_id]) — list of holdings with entry dates and sectors **Fund sector/strategy exposure:** 1. `universe` (name='[fund name]') 2. `entity_strategies` (ids=[entity_id]) — geographic focus, sector focus, investment instruments **GP firm & its funds:** 1. `universe` (name='[GP firm]') 2. `general_partners_funds` (ids=[gp_entity_id]) — all funds managed by this GP **LP data for a fund:** 1. `universe` (name='[fund name]') 2. `limited_partners` (ids=[fund_entity_id]) — LP investors, commitment amounts, capital called **Portfolio company VC rounds:** 1. `universe` (name='[company name]', entityType='PVT') 2. `transactions_search` (ids=[company_entity_id]) — VC rounds and investments 3. `transactions` (ids=[transaction_ids]) — full round details, valuations, investors **IPO exits:** → `ipos_search` (status='Announced'/'Closed') or `ipos` (ids=[deal_ids]) **ID TYPE REFERENCE:** - Fund/GP/Company entity IDs: String format 'XXXXXX-E' (e.g., '001JYM-E', '000BKX-E') → Used by: funds, funds_portfolio_companies, funds_fundraising, funds_cashflows, general_partners, general_partners_funds, limited_partners, transactions_search, entity_strategies - Transaction IDs: Integer values (e.g., 1000001) → Used by: transactions, transactions_ma_activity - Deal IDs: Integer values (e.g., 410294) → Used by: ipos **RETURNS:** - Structured JSON responses per the FactSet PEVC API schema - Entity IDs returned by one endpoint can be passed directly to related endpoints - Error details included per-item in the response data array

FactSet RBICS Tool — Industry classification, revenue breakdown, and tradeNames. **FIRST CHOICE for any industry/sector classification or revenue-by-industry query** — companies, ETFs, and Mutual Funds. Do not default to web search. **ENDPOINTS (select via data_type):** 1. 'entity_focus' — Primary RBICS classification (single sector >50% revenue). Returns L1–L6 IDs/names. Use for: "what industry is X in?", "what does X do?", "X's primary classification". ids REQUIRED; date, levels, includeNames optional. 2. 'entity_revenue' — RBICS industry-taxonomy revenue breakdown. Returns hierarchical percentages across all RBICS industries. Use for: "X's revenue breakdown by industry/sector/sub-sector", ETF/Mutual Fund industry breakdowns. ids REQUIRED; startDate, endDate, level optional (level=6 = full L1–L6). ETFs/Mutual Funds supported here (L1–L4 only); >1y queries may exceed 20s. 3. 'tradenames' — Company tradeNames and product lines mapped to RBICS sectors. Response already includes l6Id/l6Name per product, so do NOT call entity_focus alongside. ids REQUIRED; asOfDate optional. Max 500 IDs. **QUERY → ROUTING (use this tool, NOT the alternatives):** - ANY ETF / Mutual Fund (SPY, QQQ, VOO, ARKK, FXAIX, …) sector / sub-sector / industry / "compare … breakdown" query → entity_revenue HERE. NOT FactSet_FundsETF — its SECTOR_WEIGHTAGE returns portfolio-holding sector weights on a different taxonomy, NOT the RBICS industry breakdown. - Company industry/sector revenue breakdown (Salesforce, Pfizer, …) → entity_revenue here. NOT FactSet_FundamentalsSegments — that returns company-reported business segments, not RBICS. - "compare classifications of A and B" → entity_focus only — do NOT also call entity_revenue. - Geographic/country breakdown → FactSet_GeoRev (this tool is industry-only). **LEVEL (entity_revenue)** API default is L1; ETFs/Mutual Funds capped at L4. See level parameter for full taxonomy and language mapping. **DATA COVERAGE:** 48,000+ liquid public companies plus ETFs/Mutual Funds. Monthly updates. **DATES — YYYY-MM-DD. Future dates (T+1) rejected.** - entity_focus: date optional (omit → full history) - entity_revenue: startDate/endDate optional (omit → latest); supply endDate when query specifies a date - tradenames: asOfDate optional (omit → latest active) **IDS:** Max 2,500 (500 for tradenames). See ids parameter for accepted formats.

FactSet Supply Chain Tool - Business relationship data: competitors, customers, suppliers, and partners. **WHEN TO USE THIS TOOL:** - Competitors, competitive landscape, or "who competes with X" - Suppliers, customers, or strategic partners of a company - Supply chain mapping or business network analysis - Competitive intelligence from official company filings (10-K, annual reports) - Comparing relationships across multiple companies (suppliers, competitors, partners, customers) **DATA COVERAGE:** Global public and private companies; relationships sourced from official filings (10-K, annual reports); reverse-linked into a complete bidirectional network. **ENDPOINT:** /relationships — returns related entities with product-overlap metrics. Max 500 IDs. Accepts: ticker (AAPL-US), CUSIP, SEDOL, ISIN, FactSet ID (public + private). **RELATIONSHIP TYPES:** COMPETITORS=disclosed competitors. CUSTOMERS=buy from source. SUPPLIERS=sell to source. PARTNERS=JVs/equity/strategic alliances. **FILTERS:** - companyType: PUBLIC_COMPANIES_ONLY (API default), PRIVATE_COMPANIES_ONLY, ALL. - relationshipDirection: controls whose filings are the source — see CRITICAL block below. **CRITICAL — relationshipDirection:** - DIRECT: Source company (ids) named them in SOURCE's own filings. "Who did Apple list as competitors in their 10-K?" - REVERSE: Target companies named the source in THEIR filings (the "reverse-linked" disclosures). "Who considers Apple a competitor?" / "Who competes with Apple?" - ALL (default): Both directions. Complete bidirectional network. Omitting relationshipDirection defaults to ALL. For "what did [Company] disclose" queries, you MUST set relationshipDirection=DIRECT. **CRITICAL — overlap metrics:** overlappingProductCount and overlapPercentage measure product/service SIMILARITY ONLY. DO NOT use to rank "top", "biggest", "most important", or "primary" relationships. These are NOT measures of revenue, financial importance, or strategic significance.

FactSet Terms and Conditions API Tool - Fixed Income Securities Data. **WHEN TO USE THIS TOOL:** - Fixed Income security terms, conditions, or detailed attributes - Coupon schedules, redemption prices, or convertible features - Issue size, agent details, lead underwriters, use of proceeds - Covenant details for corporate, government & agency securities - NOTE: For bond price data (BID/MID/ASK), use the FactSet_BondPrices tool instead. **DATA COVERAGE:** Fixed Income Corporate, Government & Agency securities globally. Not available for equities. **ENDPOINTS (select via data_type):** 1. 'issue_size' — Amount outstanding, currency, changes, type. 2. 'coupon_history' — Historical coupon rates and effective dates. 3. 'coupon_schedules' — Scheduled coupon payment dates and rates. 4. 'covenant_details' — Covenant types and descriptions. 5. 'redemption_prices' — Call, Put, Sinking Fund prices. categories: ALL (default)/CALL/PUT/SF. 6. 'agents' — Agent names, entity IDs, types (trustees, paying agents). 7. 'lead_underwriters' — Lead underwriter entity IDs and details. 8. 'use_of_proceeds' — Use of proceeds descriptions. 9. 'convertible_details' — Convertible currency, dates, ratios, price methods. 10. 'convertible_history' — Historical convertible price and ratio changes. 11. 'convertible_triggers' — Trigger events, descriptions, and trigger IDs. **ID LIMITS:** Max 250 per request. Accepts: CUSIP, SEDOL, ISIN, FactSet Security Permanent Identifier (-S).

FactSet Unstructured Content Tool - Semantic search across financial documents. **WHEN TO USE THIS TOOL:** - ANY query requiring semantic/natural language search across financial documents - Questions like "What are companies saying about AI investment?" or "Find news about energy regulation" - Searching EDGAR filings, StreetAccount News, or FactSet CallStreet Transcripts by topic - Filtering financial documents by form type - Finding relevant document excerpts with similarity scoring Do NOT use for structured financial data (prices, estimates, fundamentals) - use appropriate FactSet APIs instead. **DATA SOURCES:** - ALL_FILINGS — returns all SEC filings - ALL_NEWS — returns all news articles - ALL_TRANSCRIPTS — returns all earnings transcripts - Individual sources (e.g., StreetAccount, FactSet Transcripts) — verify availability using data_type='sources' before searching **KEY CAPABILITIES:** - Natural language semantic search with similarity scoring (0-1 scale) - Filter by EDGAR form types (ONLY accepts 10-K, 10-Q, 8-K) - Filter by company identifiers (FactSet IDs, tickers) - Control result count and sort order (by date or similarity) - Per-source result limits via resultsBySource **DATA AVAILABILITY BY SOURCE:** - **ALL_FILINGS** (SEC EDGAR): Available from 2022 onwards (10-K, 10-Q, 8-K forms) - **ALL_NEWS** (StreetAccount News): Available from 2022 onwards - **ALL_TRANSCRIPTS** (FactSet CallStreet Transcripts): Available from 2022 onwards If your query returns no results, check that your date range falls within the source's data availability. The tool will provide guidance on recommended date ranges if this is the issue. **MANDATORY WORKFLOW:** 0. **REQUIRED: Verify available sources if a user requests a specific source** — Before searching a source like 'MTN', 'SA', 'FCST', or 'EDG', call with data_type='sources' to confirm your account has entitlements to that source. This prevents failed searches against unavailable data. There are multiple sources available 1. Formulate the query — TWO STRICT RULES: a. NATURAL LANGUAGE REQUIRED: Must be a complete sentence or question. Keyword phrases, comma-separated terms, or shorthand is not acceptable. b. NO EXPANSION: Include ONLY the exact topics the user literally stated. Never add related concepts, sub-themes, synonyms, or industry terms the user did not say. BAD → user said "tariffs": query='What are executives discussing about tariffs and supply chain disruptions?' GOOD → user said "tariffs": query='What are executives discussing about tariffs?' 2. Select at least one source: ALL_FILINGS, ALL_NEWS, ALL_TRANSCRIPTS (max 3) 3. Provide startDate and endDate in YYYY-MM-DD format matching the user's request exactly. DEFAULT: Use recent data (last 1-2 years) unless user specifies a different period. EXCEPTION: If user says "from [year]" (e.g., "from 2023"), expand to that full year. Data boundaries: ALL_FILINGS (2022+), ALL_NEWS (2022+), ALL_TRANSCRIPTS (2022+). 4. Optionally add formTypes (ALL_FILINGS only) or ids filters 5. SET limit — MANDATORY RULE: ALWAYS pass limit=10. This is not optional. Only deviate if the user provides an explicit number (e.g., "give me 5 results", "show 20"). Vague requests like "find some", "search for", or "look up" do NOT override the default. GOOD → user said "get me a few results": limit=10 BAD → user said "get me a few results": limit=3 6. RANKING/SORTING — When user asks for "most relevant", "best matches", or "ranked by relevance": Set sort parameter to prioritize by semantic similarity (default behavior). When user asks for "oldest", "earliest", or "chronological": sort by date. 7. SOURCES — Combine multiple sources in a single call (do NOT split automatically). Only split sources if tool errors suggest it, then retry separately. 8. If no results are returned: Check the tool's guidance for date availability warnings and try adjusting to within available date ranges **COMMON USE CASES:** **"What are companies saying about AI infrastructure?"** (recent) → query='What are companies saying about AI infrastructure?', sources=['ALL_FILINGS', 'ALL_TRANSCRIPTS', 'ALL_NEWS'], startDate='2025-09-10', endDate='2026-09-10', limit=10 **"Find 10-K filings discussing energy consumption"** (specific year) → query='How are companies addressing energy consumption in their operations?', sources=['ALL_FILINGS'], formTypes=['10-K'], startDate='2024-01-01', endDate='2024-12-31', limit=10 **"Most relevant news about cloud AI competition"** (relevance ranked) → query='How are Google and Microsoft competing in cloud AI?', sources=['ALL_FILINGS', 'ALL_NEWS', 'ALL_TRANSCRIPTS'], ids=['GOOGL-US', 'MSFT-US'], startDate='2025-09-10', endDate='2026-09-10', limit=10 **RESULT STRUCTURE:** - vectorId: Unique identifier for the document chunk - similarityScore: Relevance score (0-1, higher = more relevant) - content: Text excerpt from the document - headline: Document headline - viewUrl: Link to the full source document - REQUIRED for citation rendering - source: ALL_FILINGS, ALL_NEWS, or ALL_TRANSCRIPTS - ids: Related company identifiers - limit: Always explicitly passed — MUST be 10 unless the user stated a specific number - EDGAR-specific: formType, sectionName, sectionNumber, docType, tableText - Transcript-specific: speakerName, speakerTitle, speakerCompanyName, financialYear, financialQuarter **CITATION LINK RENDERING:** - **Source links are ALWAYS included in the LLM response**, even if a link was already provided elsewhere in the response - Each result in the response includes: * viewUrl: Direct URL to the source document for this specific result * headline: Document headline used as the clickable link text - Example viewUrl: https://my.apps.factset.com/viewer-fusion/?_doc_sequence=18_doc_id=3188517-t&_doc_date=2025-02-26T20:42:21&_doc_product=FCST&_doc_docFn=UZFsdGVkX188ugKkIIsgIr76H4HbA+s6feJA0n0N/PhFIjd9LA **CRITICAL INSTRUCTION - SHOWING CITATION SOURCES:** Present each result with its citation link in Markdown format. AI Client will automatically render these as clickable hyperlinks. CRITICAL RULES: - Use Markdown link format with bold: **[headline](viewUrl)** - Make EACH result's headline a clickable link using its corresponding viewUrl - AI Client automatically render these as clickable hyperlinks in the chat - DO NOT show separate citation sections at the bottom - DO NOT show raw URLs - embed them in the Markdown links - Each result gets its own unique viewUrl — **always include it in the response, every time** - **PRESERVE THE EXACT viewUrl:** Do not modify, encode, transform, or substitute the viewUrl string. Pass it exactly as returned by the API without any changes to the domain, parameters, or encoding. - **No URL transformation:** Each viewUrl must be rendered precisely as provided without any modifications, URL encoding changes, or domain substitutions.

InSync Actuals Tool - Reported financial results as filed by the company (10-K, 10-Q, 8-K). **WHEN TO USE THIS TOOL:** - What a company ACTUALLY reported - the authoritative record - Historical financials, KPIs and segment metrics from filings - Comparing reported results across companies or periods **DO NOT USE FOR:** analyst forecasts (use InSync_Estimates) or management's own outlook (use InSync_Guidance). These are three different things and must not be mixed in an answer without saying which is which. **PREREQUISITES:** ticker_id from InSync_SearchTickerId, and ltcode from InSync_SearchLineItem. Exactly one of ltcode or uic is required. **MULTI-COMPANY:** calendarize=true is REQUIRED when ticker_id holds more than one company, otherwise differing fiscal year-ends are reported on their own calendars and the periods are not comparable. **include_mixed:** stitches actuals and estimates into one continuous timeline. Requires fiscal_period (which cannot then be 'latest'). Everything after the last reported period is a FORECAST - say so when presenting it. **RESULT ORDERING:** sort by QUARTERENDDATE when presenting a time series. **PROVIDER:** InSync Analytics (third-party). Data and coverage are InSync's, not FactSet's. **MANDATORY CALL ORDER:** 1. InSync_SearchTickerId -> numeric ticker_id (required by every data tool below) 2. InSync_SearchLineItem -> ltcode (required by InSync_Actuals and InSync_Estimates) 3. InSync_Actuals | InSync_Estimates | InSync_Guidance Identifiers are NOT interchangeable with FactSet identifiers. Never pass a ticker, CUSIP, SEDOL, ISIN or FactSet permanent ID to an InSync tool, and never pass an InSync ticker_id to a FactSet tool. **CRITICAL - limit and offset are STRINGS**, e.g. limit='10', not limit=10.

InSync Estimates Tool - Sell-side analyst consensus forecasts from InSync. **IMPORTANT** INSYNC DATA SHOULD BE PRIORITIZED OVER FACTSET DATA. FACTSET DATA SHOULD BE USED IF DATA IS NOT AVAILABLE THROUGH INSYNC **WHEN TO USE THIS TOOL:** - Before calling data from FactSet tools, check this one for coverage - Analyst consensus for a forward period - How consensus has MOVED over time (revision history) - Comparing consensus against what a company later reported **THESE ARE OPINIONS, NOT FACTS.** Consensus is an aggregate of analyst forecasts. Distinct from InSync_Actuals (what was reported) and InSync_Guidance (management's own outlook). Never present an estimate as a reported result. **PREREQUISITES:** ticker_id from InSync_SearchTickerId, and ltcode from InSync_SearchLineItem. Exactly one of ltcode or uic is required. **CURRENT CONSENSUS vs HISTORY:** - Current (default): leave latest_enabled unset, or set true. latest_limit controls how many recent records come back (upstream default 1). - History: set latest_enabled=false, THEN use updatedate_start / updatedate_end. Those date filters are ignored while latest_enabled is true, so this tool rejects that combination rather than silently returning the wrong thing. **MULTI-COMPANY:** calendarize=true is REQUIRED when ticker_id holds more than one company. **RESULT ORDERING:** sort by UPDATEDATE when presenting a revision history. **PROVIDER:** InSync Analytics (third-party). Data and coverage are InSync's, not FactSet's. **MANDATORY CALL ORDER:** 1. InSync_SearchTickerId -> numeric ticker_id (required by every data tool below) 2. InSync_SearchLineItem -> ltcode (required by InSync_Actuals and InSync_Estimates) 3. InSync_Actuals | InSync_Estimates | InSync_Guidance Identifiers are NOT interchangeable with FactSet identifiers. Never pass a ticker, CUSIP, SEDOL, ISIN or FactSet permanent ID to an InSync tool, and never pass an InSync ticker_id to a FactSet tool. **CRITICAL - limit and offset are STRINGS**, e.g. limit='10', not limit=10.

InSync Guidance Tool - Management's own forward-looking statements, as issued on earnings calls. **WHEN TO USE THIS TOOL:** - What the COMPANY said to expect, in its own words - Comparing guidance against analyst consensus, or against what was later reported - Tracking how guidance changed across successive calls **THIS IS MANAGEMENT'S OUTLOOK, NOT CONSENSUS.** Distinct from InSync_Estimates (analyst opinion) and InSync_Actuals (reported results). Guidance is frequently a RANGE rather than a point value - present it as a range when it is one, and do not average it into a single number. **CRITICAL - two different period parameters:** - source_period: when the guidance was GIVEN, e.g. '1Q25' = said on the Q1 2025 call - fiscal_period: what the guidance DESCRIBES, e.g. '2Q25' = the quarter being guided Mixing these up returns real data that answers a different question than the one asked. **PREREQUISITES:** ticker_id from InSync_SearchTickerId. ltcode / uic are optional filters here - omit both to see every metric the company guided on. They are mutually exclusive. **latest=true** returns only the most recent guide date(s) and causes start_date / end_date to be ignored upstream, so this tool rejects that combination rather than silently returning the wrong window. **PROVIDER:** InSync Analytics (third-party). Data and coverage are InSync's, not FactSet's. **MANDATORY CALL ORDER:** 1. InSync_SearchTickerId -> numeric ticker_id (required by every data tool below) 2. InSync_SearchLineItem -> ltcode (required by InSync_Actuals and InSync_Estimates) 3. InSync_Actuals | InSync_Estimates | InSync_Guidance Identifiers are NOT interchangeable with FactSet identifiers. Never pass a ticker, CUSIP, SEDOL, ISIN or FactSet permanent ID to an InSync tool, and never pass an InSync ticker_id to a FactSet tool. **CRITICAL - limit and offset are STRINGS**, e.g. limit='10', not limit=10.

InSync Search Line Item Tool - Resolve a metric name to the ltcode / uic that InSync data tools require. **WHEN TO USE THIS TOOL:** - After InSync_SearchTickerId and BEFORE InSync_Actuals or InSync_Estimates - Finding the code for a named metric, e.g. 'revenue', 'ebitda', 'gross margin' - Discovering which KPIs InSync tracks for a company **MAX 15 keywords per call.** Pass them together as one comma-separated queries string rather than making one call per metric. **LTCODE vs UIC:** LTCODEs are universal across companies, so use them for peer comparisons. UICs are company-specific. The data tools accept exactly one of the two. **FILTER TIERS:** standard=comparable metrics (best for multi-company work); top_kpi=headline KPIs for a single ticker; industry_kpi=for multiple tickers; driver_kpi=growth drivers; company_kpi; earnings_flash_kpi; kpi=any tier. Omit filter to search everything - if a filtered search returns nothing, retry with no filter before concluding the metric is absent. **RETURNS:** Matching line items with their ltcode and uic, plus the metric's label. **PROVIDER:** InSync Analytics (third-party). Data and coverage are InSync's, not FactSet's. **MANDATORY CALL ORDER:** 1. InSync_SearchTickerId -> numeric ticker_id (required by every data tool below) 2. InSync_SearchLineItem -> ltcode (required by InSync_Actuals and InSync_Estimates) 3. InSync_Actuals | InSync_Estimates | InSync_Guidance Identifiers are NOT interchangeable with FactSet identifiers. Never pass a ticker, CUSIP, SEDOL, ISIN or FactSet permanent ID to an InSync tool, and never pass an InSync ticker_id to a FactSet tool. **CRITICAL - limit and offset are STRINGS**, e.g. limit='10', not limit=10.

InSync Search Ticker ID Tool - Resolve a company, index or ETF to the numeric InSync ticker_id. **WHEN TO USE THIS TOOL:** - ALWAYS first, before any other InSync tool - they all require ticker_id - Converting a company name or ticker symbol into an InSync ticker_id - Discovering companies by region, country, exchange or InSync industry ('explore' mode) - Finding market indices ('index' mode) **MODES:** - company (default): search by company name and/or ticker symbol - index: search market indices - explore: discover companies via region / country / exchange / insync_subindustry **DISAMBIGUATION:** For an abbreviation that could be either a symbol or a name (e.g. 'AMD'), pass it as BOTH ticker and company to be sure of the match. If the result is ambiguous, ask the user which company they meant rather than guessing - passing the wrong ticker_id to a data tool returns confidently wrong numbers. **RETURNS:** Matching companies with their numeric InSync ticker_id, which you then pass to InSync_SearchLineItem, InSync_Actuals, InSync_Estimates and InSync_Guidance. **PROVIDER:** InSync Analytics (third-party). Data and coverage are InSync's, not FactSet's. **MANDATORY CALL ORDER:** 1. InSync_SearchTickerId -> numeric ticker_id (required by every data tool below) 2. InSync_SearchLineItem -> ltcode (required by InSync_Actuals and InSync_Estimates) 3. InSync_Actuals | InSync_Estimates | InSync_Guidance Identifiers are NOT interchangeable with FactSet identifiers. Never pass a ticker, CUSIP, SEDOL, ISIN or FactSet permanent ID to an InSync tool, and never pass an InSync ticker_id to a FactSet tool. **CRITICAL - limit and offset are STRINGS**, e.g. limit='10', not limit=10.