Exposes SEC insider trading data, including Form 4, Form 144, 13F-HR, and congressional STOCK Act trades, to AI assistants.
Also known as theodor90/form4api-mcp, form4api-mcp
- Stars
- 2
- Tools
- 40
- Downloads, 30 days
- —
- Latest version
- 1.18.0
- Last commit
- Nov 14, 2026
- License
- MIT
What is Form4API MCP?
Form4API exposes SEC insider trading data, including Form 4, Form 144, 13F-HR, and congressional STOCK Act trades, to AI assistants. It exposes 40 tools, such as get_transactions, get_recent_filings, get_filing.
Summary generated from the project's README
About Form4API MCP server
Form4API provides access to SEC insider trading data for developers using MCP-compatible AI assistants like Claude Desktop, Cursor, and Windsurf. The server connects to the Form4API REST API to deliver specialized financial intelligence, including Form 4 filings, Form 144 intent-to-sell notices, and institutional 13F-HR holdings. It is designed to help researchers identify patterns in insider activity and congressional STOCK Act trades.
Users can integrate the server by adding it to their client configuration using npx and a required FORM4API_KEY. The system distinguishes between discretionary trades and those made under 10b5-1 pre-scheduled plans to ensure data accuracy. It also features a unique capability to join insider transactions with institutional ownership context in a single response. While a free plan allows for 500 requests per day, advanced tools like cluster signals, sentiment scores, and congressional convergence require higher-tier plans.
What you can do with it
- 1Monitor insider convictionAnalyze recent Form 4 filings to determine if insiders are buying or selling a specific ticker.
- 2Identify early selling signalsDetect potential sales by tracking Form 144 filings before they appear as Form 4 transactions.
- 3Track congressional trading activityFilter and research periodic transaction reports from members of the U.S. House of Representatives.
- 4Detect insider and political convergenceFind stocks where both corporate insiders and members of Congress have recently made purchases.
Written from the project's README. Every use case is checked against a quote from it.
How to install Form4API MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Form4API MCP in Claude Code
$ claude mcp add --env 'FORM4API_KEY=<YOUR_FORM4API_KEY>' form4api-mcp -- npx -y form4api-mcpInstall Form4API MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"form4api-mcp": {
"command": "npx",
"args": [
"-y",
"form4api-mcp"
],
"env": {
"FORM4API_KEY": "<YOUR_FORM4API_KEY>"
}
}
}
}Install Form4API MCP in VS Code
.vscode/mcp.json
{
"servers": {
"form4api-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"form4api-mcp"
],
"env": {
"FORM4API_KEY": "<YOUR_FORM4API_KEY>"
}
}
}
}Install Form4API MCP in VS Code CLI
$ code --add-mcp '{"name":"form4api-mcp","type":"stdio","command":"npx","args":["-y","form4api-mcp"],"env":{"FORM4API_KEY":"<YOUR_FORM4API_KEY>"}}'Install Form4API MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"form4api-mcp": {
"command": "npx",
"args": [
"-y",
"form4api-mcp"
],
"env": {
"FORM4API_KEY": "<YOUR_FORM4API_KEY>"
}
}
}
}Install Form4API MCP in Codex
$ codex mcp add form4api-mcp --env 'FORM4API_KEY=<YOUR_FORM4API_KEY>' -- npx -y form4api-mcpInstall Form4API MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.form4api-mcp] command = "npx" args = ["-y", "form4api-mcp"] [mcp_servers.form4api-mcp.env] FORM4API_KEY = "<YOUR_FORM4API_KEY>"
Install Form4API MCP in Gemini CLI
$ gemini mcp add -s user -e 'FORM4API_KEY=<YOUR_FORM4API_KEY>' form4api-mcp npx -- -y form4api-mcpInstall Form4API MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"form4api-mcp": {
"command": "npx",
"args": [
"-y",
"form4api-mcp"
],
"env": {
"FORM4API_KEY": "<YOUR_FORM4API_KEY>"
}
}
}
}Started by HostedMCP on 16 Nov 2026
We ran this command for form4api-mcp 1.18.0 in an isolated container with no network and no keys. It answered and listed 40 tools.
Form4API MCP tools 40
Read from the server with tools/listget_transactionsRead-onlySearch SEC Form 4 insider transactions with rich filters: ticker, insider CIK, transaction code/category, date range, a dollar floor (min_value), institutional-ownership trend, and (Pro+) upper trade-size bounds plus post-trade return screening (1d/1w/1m/3m/6m, returns as fractions e.g. 0.05 = +5%). Returns transaction-level rows — shares, price, total value, transaction code, 10b5-1 flag, insider role flags. Use this for filtered or historical search across many companies/insiders; use get_recent_filings for an unfiltered live feed instead, or get_insider_transactions/get_company_insiders when you already have a specific insider or company. Free plan, including min_value; max_value/min_shares/max_shares/min_return_*/max_return_*/has_returns require Pro. Paginated, max 100/page, and paging depth is plan-limited (Free 20 pages, Starter 100, Pro+ unlimited) — for a bulk pull use the REST endpoint GET /v1/transactions/export (Business) instead of paging.
get_recent_filingsRead-onlyLive feed of the newest SEC Form 4 filings, sorted most-recent-first, optionally filtered to one ticker. New filings typically appear within ~60 seconds of SEC publication. Use this to check "what just happened" rather than get_transactions (built for filtered/historical search across date ranges and codes). Returns per-filing accession number, filed/period-of-report dates, company + insider identity, and transaction count; pass the accession number to get_filing for full detail. Free plan. Paginated, max 100/page.
get_filingRead-onlyFetch one Form 4 filing by its exact SEC accession number (format NNNNNNNNNN-YY-NNNNNN, e.g. 0000320193-26-000001). Use this once you already have an accession number from get_recent_filings or get_transactions. Returns filed/period-of-report dates, company and insider identity, and transaction count for that filing. Free plan.
get_insider_profileRead-onlyIdentity card for one insider by CIK (SEC's numeric filer identifier): name, all known titles, and director/officer/10%-owner role flags. Returns NO transactions — when the user asks what an insider bought, sold or filed, use get_insider_transactions (raw rows) or get_insider_career_summary (aggregated rollup) instead. If you only have a name, resolve it to a CIK first with search. Free plan.
get_insider_transactionsRead-onlyAll Form 4 transactions filed by one insider (by CIK), filterable by ticker, transaction code, date range, and 10b5-1 exclusion. Use this once you have an insider's CIK on hand; get_transactions with insider_cik= gives the same rows alongside its broader filter set, while get_insider_career_summary returns a pre-aggregated rollup instead of raw rows. Free plan. Paginated, max 100/page.
get_company_overviewRead-onlyCompany profile for a single ticker — name, CIK, SIC sector/description, state of incorporation, website, total Form 4 filing count, and active insider count. Use this for company identity/metadata; use get_company_insiders to list who is filing, or get_transactions with ticker= for their trade history. Free plan.
get_company_insidersRead-onlyFull roster of insiders who have ever filed a Form 4 for a given ticker — name, CIK, titles, director/officer/10%-owner flags, last-filed date, and total transaction count per insider. Use this to enumerate a company's insiders (e.g. before pulling each one's career summary); use get_transactions with ticker= for the underlying trade history itself. Free plan. Paginated, max 100/page.
get_signalsRead-onlyCluster buy/sell signals — multiple insiders at the same company trading in the same direction within a short window, a stronger conviction signal than any single trade. Excludes 10b5-1 plan trades automatically by construction. Returns signal type, detection date, buyer/seller counts, total $ value, and the underlying transactions. Use this to FIND which companies have signals; when the user already names a company and asks WHY its signal fired or which insiders were in the cluster, call explain_signal with that ticker directly instead. Use get_transactions instead for raw, unaggregated trade search. Requires Business plan (a 402 upgrade_required response is returned otherwise). Paginated, max 50/page.
get_form144Read-onlyForm 144 notice-of-proposed-sale filings — insiders disclose intent to sell roughly 2 days before the matching Form 4 sale lands, so this is an early-warning signal, especially discretionary (non-10b5-1) notices. Filter by ticker, insider name (partial match), date range, or exclude_10b5. Cross-reference with get_transactions/get_insider_transactions to see whether the intent was actually executed. Requires Business plan. Paginated, max 100/page.
get_holdingsRead-onlyInstitutional 13F-HR positions — filter by ticker (CUSIP resolved to ticker automatically), CUSIP, manager CIK, quarter, or minimum position value. Answers "who owns NVDA" or "which managers hold AAPL this quarter"; pair with get_managers to look up a manager's identity/AUM, or get_transactions to cross-reference insider activity at the same company. Requires Business plan. Paginated, max 100/page.
get_managersRead-onlyBrowse the institutional manager index — each manager with their latest 13F-HR filing date and AUM. Filter by name (partial match — "Berkshire" returns Berkshire Hathaway) or minimum AUM. Pair with get_holdings to see what a given manager owns. Requires Business plan. Paginated, max 100/page.
get_sentimentRead-onlyMonthly insider-sentiment score for a ticker (MSPR-style, -100 to +100; positive = net buying conviction), one point per month, defaulting to roughly the last 12 months. Automatically excludes 10b5-1 plan trades so the score reflects discretionary conviction, not pre-scheduled dispositions. Use get_transactions or get_signals for the trade-level detail behind a given month. Requires Business plan.
| get_transactions | Search SEC Form 4 insider transactions with rich filters: ticker, insider CIK, transaction code/category, date range, a dollar floor (min_value), institutional-ownership trend, and (Pro+) upper trade-size bounds plus post-trade return screening (1d/1w/1m/3m/6m, returns as fractions e.g. 0.05 = +5%). Returns transaction-level rows — shares, price, total value, transaction code, 10b5-1 flag, insider role flags. Use this for filtered or historical search across many companies/insiders; use get_recent_filings for an unfiltered live feed instead, or get_insider_transactions/get_company_insiders when you already have a specific insider or company. Free plan, including min_value; max_value/min_shares/max_shares/min_return_*/max_return_*/has_returns require Pro. Paginated, max 100/page, and paging depth is plan-limited (Free 20 pages, Starter 100, Pro+ unlimited) — for a bulk pull use the REST endpoint GET /v1/transactions/export (Business) instead of paging. | Read-only |
| get_recent_filings | Live feed of the newest SEC Form 4 filings, sorted most-recent-first, optionally filtered to one ticker. New filings typically appear within ~60 seconds of SEC publication. Use this to check "what just happened" rather than get_transactions (built for filtered/historical search across date ranges and codes). Returns per-filing accession number, filed/period-of-report dates, company + insider identity, and transaction count; pass the accession number to get_filing for full detail. Free plan. Paginated, max 100/page. | Read-only |
| get_filing | Fetch one Form 4 filing by its exact SEC accession number (format NNNNNNNNNN-YY-NNNNNN, e.g. 0000320193-26-000001). Use this once you already have an accession number from get_recent_filings or get_transactions. Returns filed/period-of-report dates, company and insider identity, and transaction count for that filing. Free plan. | Read-only |
| get_insider_profile | Identity card for one insider by CIK (SEC's numeric filer identifier): name, all known titles, and director/officer/10%-owner role flags. Returns NO transactions — when the user asks what an insider bought, sold or filed, use get_insider_transactions (raw rows) or get_insider_career_summary (aggregated rollup) instead. If you only have a name, resolve it to a CIK first with search. Free plan. | Read-only |
| get_insider_transactions | All Form 4 transactions filed by one insider (by CIK), filterable by ticker, transaction code, date range, and 10b5-1 exclusion. Use this once you have an insider's CIK on hand; get_transactions with insider_cik= gives the same rows alongside its broader filter set, while get_insider_career_summary returns a pre-aggregated rollup instead of raw rows. Free plan. Paginated, max 100/page. | Read-only |
| get_company_overview | Company profile for a single ticker — name, CIK, SIC sector/description, state of incorporation, website, total Form 4 filing count, and active insider count. Use this for company identity/metadata; use get_company_insiders to list who is filing, or get_transactions with ticker= for their trade history. Free plan. | Read-only |
| get_company_insiders | Full roster of insiders who have ever filed a Form 4 for a given ticker — name, CIK, titles, director/officer/10%-owner flags, last-filed date, and total transaction count per insider. Use this to enumerate a company's insiders (e.g. before pulling each one's career summary); use get_transactions with ticker= for the underlying trade history itself. Free plan. Paginated, max 100/page. | Read-only |
| get_signals | Cluster buy/sell signals — multiple insiders at the same company trading in the same direction within a short window, a stronger conviction signal than any single trade. Excludes 10b5-1 plan trades automatically by construction. Returns signal type, detection date, buyer/seller counts, total $ value, and the underlying transactions. Use this to FIND which companies have signals; when the user already names a company and asks WHY its signal fired or which insiders were in the cluster, call explain_signal with that ticker directly instead. Use get_transactions instead for raw, unaggregated trade search. Requires Business plan (a 402 upgrade_required response is returned otherwise). Paginated, max 50/page. | Read-only |
| get_form144 | Form 144 notice-of-proposed-sale filings — insiders disclose intent to sell roughly 2 days before the matching Form 4 sale lands, so this is an early-warning signal, especially discretionary (non-10b5-1) notices. Filter by ticker, insider name (partial match), date range, or exclude_10b5. Cross-reference with get_transactions/get_insider_transactions to see whether the intent was actually executed. Requires Business plan. Paginated, max 100/page. | Read-only |
| get_holdings | Institutional 13F-HR positions — filter by ticker (CUSIP resolved to ticker automatically), CUSIP, manager CIK, quarter, or minimum position value. Answers "who owns NVDA" or "which managers hold AAPL this quarter"; pair with get_managers to look up a manager's identity/AUM, or get_transactions to cross-reference insider activity at the same company. Requires Business plan. Paginated, max 100/page. | Read-only |
| get_managers | Browse the institutional manager index — each manager with their latest 13F-HR filing date and AUM. Filter by name (partial match — "Berkshire" returns Berkshire Hathaway) or minimum AUM. Pair with get_holdings to see what a given manager owns. Requires Business plan. Paginated, max 100/page. | Read-only |
| get_sentiment | Monthly insider-sentiment score for a ticker (MSPR-style, -100 to +100; positive = net buying conviction), one point per month, defaulting to roughly the last 12 months. Automatically excludes 10b5-1 plan trades so the score reflects discretionary conviction, not pre-scheduled dispositions. Use get_transactions or get_signals for the trade-level detail behind a given month. Requires Business plan. | Read-only |
Show all 40 toolsShow fewer tools
get_insider_career_summaryRead-onlyAggregate career rollup for one insider by CIK — first/last transaction dates, total shares/value bought vs sold, top companies traded, transaction-code breakdown, 10b5-1 plan split, and median post-trade returns (stored as fractions, e.g. 0.05 = +5%). Use this instead of get_insider_transactions when you want a pre-computed summary rather than raw rows; pair with the generated get_insider_scorecard for hit-rate statistics on their discretionary buys specifically. Requires Pro plan.
check_usageRead-onlySnapshot of the authenticated API key's current usage — plan name, requests made today, daily limit, and all-time request count. Use this for a quick right-now check; use the generated get_usage_history tool for a daily trend over time, or get_key_activity for a per-request log. Free plan, works on every tier.
verify_setupRead-onlyDiagnostic check: confirms FORM4API_KEY is present and valid, the backend is reachable, and reports your current plan — returns a pass/fail/warn per check plus concrete next steps (where to get a key, how to upgrade) instead of a raw error. Run this first whenever another tool fails or returns a 401/402, to isolate whether the problem is configuration, plan, or a backend outage. No API key or parameters required.
searchRead-onlyResolve free-text input — a company name, ticker fragment, or a person's name — to a ticker or CIK. This is usually the right FIRST call when a user names a company or insider you don't already have an identifier for; feed the resulting ticker into get_company_overview/get_transactions or the CIK into get_insider_profile/get_insider_transactions. Returns two independently-ranked lists: `companies` (only ones with a resolved public ticker, ranked exact ticker match, then ticker-prefix, then name-prefix, then name-contains) and `insiders` (matched by splitting the query on whitespace so every token must match the name, e.g. "tim cook" matches SEC-style "Cook Timothy D"; ranked by total filing count, most active first). Each insider's `ticker` is a ticker associated with that insider (may be null). `q` must be 2-64 characters after trimming (400 QUERY_TOO_SHORT/QUERY_TOO_LONG). `limit` applies independently to each list, default 8, max 20. Free plan, no caching.
research_companyRead-onlyBundled insider-research context for one ticker in a single call — company profile, recent insider transactions (10b5-1 flagged), cluster buy/sell signals, monthly sentiment score, and a computed net buy/sell direction summary. Use this FIRST for any company insider-research question; it replaces 4 separate calls (get_company_overview, get_transactions, get_signals, get_sentiment) and degrades gracefully — if signals/sentiment require a plan you don't have, they come back null with an explanatory `_unavailable` note instead of failing the whole call. Company/transaction data works on Free; signals/sentiment sections require Business. Fetches all sections concurrently. Research and decision-support only, not investment advice.
explain_signalRead-onlyExplain why a signal fired: the insiders and trades counted, what was excluded, and the criteria (Business plan+). Use this when the question is WHY a named company's signal fired, or WHICH insiders and trades made up its cluster. It takes the ticker directly, so there is no need to call GET /v1/signals first; that endpoint is for finding which companies have signals. Reconstructs the full evidence behind one company's insider signal from GET /v1/signals: the detection criteria (5-day cluster window, 3-insider threshold, 90-day ratio window, 10b5-1 and derivative-row exclusion), the list of cluster buyers and sellers (each with their role and individual trades in the window), trades that were excluded from the cluster count and why (10b5-1 plan, derivative security, or superseded by amendment), and the raw buy/sell share totals behind the 90-day ratio. Use this to audit or debug a specific signal rather than to scan many companies (use GET /v1/signals for that). This is a LIVE reconstruction from current non-superseded data, computed on every request (no caching) — it can differ slightly from the originally stored signal if trades were amended afterward. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). Returns 404 COMPANY_NOT_FOUND if the ticker isn't tracked, 404 SIGNAL_NOT_FOUND if no signal exists for the given/most-recent date, or 400 INVALID_DATE if `date` isn't YYYY-MM-DD.
get_company_peersRead-onlyPeer companies in the same SIC industry, ranked by recent Form 4 activity. Returns other tracked companies in the same industry as the given ticker, for "similar companies" links. Candidates share the company's 4-digit SIC code (matchLevel "sic4"); when too few of those have Form 4 filings in the last 12 months to fill `limit`, companies sharing only the 2-digit SIC prefix (matchLevel "sic2") are added. Active peers (at least one Form 4 filing dated in the trailing 12 months, `filings12m`) come first, sic4 before sic2, each tier ordered by `filings12m` descending then ticker ascending; peers with no filings in the window are appended last only if there is still room. The company itself is excluded, as are companies without a ticker or without a usable SIC code, and any other company holding the same ticker. When several companies hold the ticker, the one GET /v1/companies/{ticker} returns is used. `basis` is "sic4", "sic2" or "mixed" according to the match levels present in `peers`, and "none" when `peers` is empty (including when the company has no SIC code, which is 200 with an empty list, not 404). `limit` defaults to 10 and is clamped to 1-25. Returns 404 NOT_FOUND if the ticker isn't tracked. Not plan-gated. Results are cached for a few hours.
get_company_returns_summaryRead-onlyWhat happened after insiders traded at a company (Pro plan+). Summarises price returns after insider open-market trades at one company, for buys and for discretionary sells, next to a site-wide baseline computed the same way. The unit is an event: one per company per filing date and side. An event's return at a horizon (1m, 3m, 6m = 21, 63, 126 trading days) is the average of that event's rows, measured from the filing-date close. Included: open-market purchases (P) and sales (S) that are not superseded, not derivative and carry no value-quality flag, filed since 2023-05-22; sells exclude 10b5-1 plan sales. An event counts toward a horizon only once that horizon has matured. `nEvents` is the number of events with a 3m return; `n1m` and `n6m` are the counts for the other horizons. Medians are null when that horizon has fewer than 10 events; `mean3m` and `hitRate3m` (share of events with a positive 3m return) are null when `nEvents` is below 10. `events` lists the individual events, newest first, only when `nEvents` is between 3 and 9, otherwise it is empty. `suppressed` is true when the median filing-date close across the company's buy and sell events is below $1, and `lowPrice` is true when that median is from $1 up to below $5; the price itself is not returned. Market comparison: each event's stock return is compared with the S&P 500 (SPY) return over the same window, from the stock's filing-date close bar to the stock's own bar at the horizon, matched by exact date; the event's excess return is the stock return minus the SPY return, computed only when both exist (events without an SPY return are left out, never filled in). `nExcess1m`, `nExcess3m` and `nExcess6m` count the events with an excess return per horizon; `excessMedian1m`, `excessMedian3m` and `excessMedian6m` are null when that count is below 10; `excessMean3m` and `beatMarket3m` (share of events whose 3m stock return exceeded SPY's) are null when `nExcess3m` is below 10. Each item in `events` carries `excess3m`. Not adjusted for company size or sector. The `baseline` block (nEvents, median3m, mean3m, hitRate3m, plus excessMedian3m, excessMean3m and beatMarket3m for buys and sells) is site-wide. All return fields are FRACTIONS, 0.05 means +5%. Requires Pro plan or higher (402 PLAN_REQUIRED on Free). Returns 404 NOT_FOUND if the ticker isn't tracked; when several companies hold the ticker, the one GET /v1/companies/{ticker} returns is used. Cached for 24 hours per ticker, and the baseline is cached for 24 hours, so `asOf` can lag the latest filings by up to a day or two.
get_congress_politicianRead-onlyGet one politician's congressional trading profile (Pro plan+). Returns one politician's profile — identity fields, total/buy/sell trade counts, most recent trade's disclosure date, their top N most-traded tickers (by trade count), and their N most recent trades (same shape as GET /v1/congress/trades). Use this for a one-call politician overview rather than paging /v1/congress/trades?politician= yourself. Accepts either a bioguide ID (e.g. "P000197") or the politician's URL slug (e.g. "nancy-pelosi") in the path, matched case-insensitively against whichever field applies. Returns 404 NOT_FOUND if neither matches. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Query runs live — no caching.
get_congress_ticker_rollupRead-onlyWhich politicians traded a ticker, with net buy/sell counts (Pro plan+). Returns every politician who has a non-superseded congressional trade in the given ticker, each with their trade/buy/sell counts, plus ticker-level totals. Optional window_days restricts to trades with a transactionDate in the trailing N days; omit for all-time. A ticker with no congress trades returns 200 with an empty politicians array and zero counts rather than 404 — there is no separate ticker/company entity in this dataset to 404 against. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Query runs live — no caching.
get_convergence_signalsRead-onlyInsider cluster-buy x congressional-purchase convergence (Pro plan+). Returns the tickers where an insider cluster-buy (InsiderSignal.IsClusterBuy) and at least one non-superseded congressional PURCHASE happened within window_days of EACH OTHER, restricted to convergences where the MORE RECENT of the pair's two dates is within a trailing lookback_days (so this surfaces CURRENT convergences, not ancient history). DEFINITION: for each result, insider.signalDate is the SignalDate of the qualifying cluster-buy signal with the most recent date (insider.insiderCount is that same signal's count — never summed or maxed across multiple signals), and congress is every non-superseded congressional purchase that paired with at least one qualifying cluster-buy (not every purchase in the window — only the ones that actually paired). firstSeen/lastSeen are the earliest/most recent dates among all qualifying insider and congress dates for that ticker. STRENGTH is documented arithmetic, NOT a black-box or predictive/ML score: strength = (distinct congressional purchasers among the qualifying legs) x (the representative signal's insiderCount) — a plain multiplication of two observed counts, nothing more. HONESTY: every congress leg always carries both amountLow and amountHigh (STOCK Act discloses ranges, never exact figures — never combined into a fabricated midpoint) and disclosureLagDays = (disclosureDate - transactionDate). Congressional trades whose dates are impossible or implausible (a dateQuality flag: transaction_after_disclosure, future_transaction_date or implausible_lag) are excluded from detection, so they never appear as legs and never create a convergence; dateQuality is therefore null on every leg returned here and disclosureLagDays is always set. Congressional trades are disclosed up to 45 days after the actual trade under the STOCK Act, so this endpoint is detection/monitoring of what insiders AND members of Congress have DISCLOSED buying, not a claim of predictive edge, alpha, or win rate — no performance numbers are computed or implied anywhere in this response. window_days and lookback_days are both caller-overridable with clamps (see each parameter's own description for the exact bounds). Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching.
get_data_qualityRead-onlyPublic data-quality, freshness and coverage metrics for the whole dataset. Returns public, keyless metrics on data freshness, ingestion latency, corpus coverage, and post-trade returns coverage — use this to check whether the dataset is current before relying on it (e.g. confirm Form 4 ingestion isn't stalled, or that price data isn't stale), not to look up any single company, insider, or transaction. Includes: most recent Form 4 processed timestamp and median/p95 filing-accepted-to-processed latency in seconds, latest price-bar date and how many days behind it is, total companies/transactions tracked plus filing counts by form type (4, 144, 13F-HR), the percentage of 13F CUSIPs resolved to a ticker, and the percentage of eligible transactions with fully computed post-trade returns. Takes no parameters. Cached for 30 minutes; no API key or plan required.
get_insider_directoryRead-onlyBrowse insiders alphabetically by surname. Returns the A-Z rail with a count per letter, plus one page of insiders under the requested letter. Omit `letter` to get the rail and totals with no rows. Names come from EDGAR surname-first ("HENNEMAN JOHN B III"), so alphabetical order is order by surname. Casing in the source is inconsistent and is not normalised here. This lists only insiders with at least 3 non-superseded transactions, capped at the 5,000 most active — the same set as the insiders sitemap shard, so the two cannot drift. To find someone outside that set, use GET /v1/insiders?name= which searches every filer. Rebuilt daily; `refreshedAt` reports when. `limit` is accepted as an alias for `per_page`. Not plan-gated. One row per FILER GROUP. A fund group files a single Form 4 listing several reporting owners — the fund, its GP, its management company — and each is a real EDGAR filer with its own CIK. Listing all of them spent about 11% of this capped surface describing the same actors more than once, so browse shows one per group and `filerGroupSize` says how many others share those exact transactions. The others are not hidden: each keeps its own profile and is still returned by GET /v1/insiders?name=.
get_insider_leaderboardRead-onlyRanked leaderboard of insiders by buy track-record (Business plan+). Returns the top insiders ranked by historical buy performance — same scored-buy methodology as GET /v1/insiders/{cik}/scorecard, applied across the whole corpus rather than one insider. Use this to discover which insiders have the best track record; use the per-insider scorecard once you have a specific CIK. Scores use absolute return (NOT market-adjusted) — a hit is a scored buy with a positive 3m (or 6m) return anchored at the filing-date close. Only discretionary open-market buys (P-code, not 10b5-1, not derivative) with a matured return are counted. Insiders with fewer than min_trades (floor 5) scored buys are excluded. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). Results are cached for 1 hour per unique parameter combination.
get_insider_scorecardRead-onlyGet insider buy track-record scorecard (Pro plan+). Returns the historical hit rate and average/median return of an insider's discretionary open-market buys (TransactionCode=P, excluding 10b5-1 plans and derivatives), plus their best and worst scored buy. Scores use absolute return (NOT market-adjusted) anchored at the filing-date close. A 'hit' is a scored buy whose 3m (or 6m) return is positive. Use this over GET /v1/insiders/{cik}/summary when you specifically want a scored track record (with a sample-sufficiency guard) rather than raw totals; use GET /v1/insiders/leaderboard (Business+) to rank many insiders by this same methodology. Score fields (hitRate3m, avgReturn3m, etc.) are null when the insider has fewer than 5 matured scored buys (sampleSufficient=false), preventing misleading statistics from small samples. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Returns 404 NOT_FOUND if the CIK isn't tracked. Computed live — no caching. Note: all return fields (HitRate3m, AvgReturn3m, MedianReturn3m, etc.) are stored as FRACTIONS — 0.05 means +5%, -0.10 means -10%.
get_key_activityRead-onlyRecent, per-request API activity log for the authenticated key. Returns the most recent HTTP requests made with the authenticated API key, most recent first, including the endpoint path, response status code, duration in milliseconds, and timestamp. Use this to debug integration issues — confirm a specific call reached the API, check for repeated 4xx/5xx responses, or spot slow requests — rather than for usage trends; for aggregate daily counts use GET /v1/keys/usage/history instead. Requires a valid X-Api-Key (401 without one). Limited to 30 requests per minute per API key, shared with the other /v1/keys/usage routes (429 KEY_USAGE_RATE_LIMIT).
get_public_statsRead-onlyPublic corpus-wide statistics — no API key required. Returns headline dataset totals: filing count, transaction count, tracked companies, institutional holdings rows, Form 144 and Form 13F-HR filing counts, the earliest filing date in the corpus, the most recent quarter's total 13F-HR reported AUM in USD, and measured ingestion latency (median/p95 seconds from SEC acceptance to our processing, trailing 7 days). Use this for corpus-wide totals (e.g. a marketing/status widget), not for per-company or per-insider data — those live under GET /v1/companies and GET /v1/insiders. For freshness and coverage-quality metrics (is ingestion stalled, is price data stale) use GET /v1/data-quality instead. Takes no parameters. No API key or plan required. Cached for ~12 hours.
get_status_historyRead-onlyMeasured uptime history for the public status page — trailing 90-day daily breakdown. Returns a daily breakdown of measured API uptime over a trailing 90-day window, computed from an internal heartbeat probe that runs every 5 minutes and performs the same DB-connectivity check as GET /health/ready. Each day in the `days` array reports the number of 5-minute slots expected to have elapsed (288 for a complete past day, pro-rated for the feature's first day and for today's partial day), how many of those slots recorded a healthy heartbeat, and the resulting uptime percentage for that day — plus an overall percentage (`overallPct`) across the whole window. `start` is the earliest date included: either the date of the very first heartbeat ever recorded, or 89 days before today once more than 90 days of history exist. Days before that are never returned. Use this to render an uptime history / status bar; for live corpus freshness use GET /v1/data-quality instead. Takes no parameters. Cached for ~5 minutes; no API key or plan required.
get_usage_historyRead-onlyDaily request counts for the authenticated key over a trailing window. Returns a daily time series of request counts for the authenticated API key over the trailing N days — one data point per calendar day (UTC). Use this to plot usage trends or check rate-limit headroom over time. For a single current-day snapshot (today's count, plan limit, reset time) use GET /v1/keys/usage instead; for a raw request-by-request log use GET /v1/keys/usage/activity. Requires a valid X-Api-Key (401 without one). Limited to 30 requests per minute per API key, shared with the other /v1/keys/usage routes (429 KEY_USAGE_RATE_LIMIT).
get_webhook_eventsRead-onlyReplay webhook delivery events since a given timestamp (default: last 24h). Returns up to 500 delivery attempts across all of the authenticated key's subscriptions since `since` (default: last 24 hours), most recent first — delivery id, subscription id, event type, attempt count, delivered-at/next-retry-at timestamps, last HTTP status code from the receiving endpoint, whether the delivery is dead (exhausted all retries), and the event payload. Use this to reconcile missed webhook deliveries (e.g. after an outage on your receiving endpoint) rather than relying solely on push delivery. `since` cannot be more than 30 days in the past. Requires a valid X-Api-Key (401 without one). `eventType` on each row is the exact name accepted by POST /v1/webhooks and sent in the X-Event-Type delivery header (e.g. "CongressTradeFiled", "ConvergenceSignal" — not dotted names). `payload` is null and `payloadRedacted` is true for any event type above your current plan (CongressTradeFiled requires Starter, ConvergenceSignal requires Pro) — delivery history outlives the plan that created it, so payloads are checked against the plan you are on now, not the plan you had when you subscribed.
health_ingestionRead-onlyIngestion-death detector — Form 4 freshness, parse-queue health, price-feed freshness. Returns live (uncached) ingestion health: Form 4 processing freshness (stale after 90 minutes during the weekday 06:00-22:00 US Eastern EDGAR activity window, counted from the later of the last filing and the 06:00 opening; outside it, 14 hours or the time since the window last closed plus 90 minutes, whichever is longer, so a normal weekend never reads as stale), parse-queue health (stale when >50 jobs are pending AND the oldest has waited >45 minutes), and daily price-feed freshness (degraded, never stale, when price bars are >5 days behind). Overall status is "healthy", "degraded" (price feed only), or "stale" (Form 4 or queue). Returns HTTP 503 when stale, 200 otherwise — safe to point an external uptime monitor at directly. No API key or plan required.
list_companiesRead-onlyList companies with a public ticker, sorted by name or total filings. Returns a single page of companies that have a tracked public ticker — for browsing or building a company picker, not for searching by name or CIK (there is no full-text search here; use GET /v1/companies/{ticker} to fetch one company by its exact ticker). Each entry includes the company's CIK, name, ticker, exchange, total filing count, and distinct insider count. There is no page parameter — this endpoint always returns the top `limit` companies by the chosen sort order. Not plan-gated.
list_congress_politiciansRead-onlyRanked rollup of politicians by congressional trade activity (Pro plan+). Returns a paginated list of politicians who have at least one non-superseded congressional trade, each with total/buy/sell counts (sells include both Sale and PartialSale; Exchange trades count only toward total) and their most recent trade's disclosure date. Ordered by total trade count descending, ties broken by most recently disclosed. Use this to discover active traders; for one politician's full profile (including their most-traded tickers and recent trades) use GET /v1/congress/politicians/{idOrSlug}. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). `limit` is accepted as an alias for `per_page`. Query runs live — no caching.
list_congress_tradesRead-onlyQuery congressional STOCK Act trades (Free+, plan-clamped disclosure window). Returns a paginated JSON list of congressional periodic-transaction-report trades, most recently DISCLOSED first, with non-superseded rows only (amended-away rows never appear). COVERAGE — HOUSE ONLY TODAY: every trade in this dataset comes from the U.S. House Clerk's PTR index. Senate eFD (efdsearch.senate.gov) returns 403 to datacenter traffic, so no Senate filings are ingested yet. chamber=Senate remains a valid filter but matches nothing and returns the response header X-Coverage-Note: chamber-not-covered, so an empty result is never ambiguous. Scanning by chamber should treat that header as "not covered", not as "no trades". PLAN-CLAMPED WINDOW: this endpoint is open to every plan, but how far back you can see is clamped on disclosureDate — Free sees only trades disclosed in the last 30 days, Starter the last 366 days, Pro/Business/Enterprise unlimited history. Passing an older disclosure_date_from than your plan allows does not extend the window — the floor always wins. Filters: ticker, politician (bioguideId, exact), party (free-text, case-insensitive exact match — not a fixed enum), chamber (House|Senate — see the coverage note above), state (2-letter code), transaction_type (purchase|sale|partial_sale|exchange), min_amount (range-aware — matches AmountLow >= value, never a fabricated midpoint), transaction_date_from/to, disclosure_date_from/to. Every row always carries BOTH amountLow and amountHigh (STOCK Act discloses ranges, never exact figures) and disclosureLagDays = (disclosureDate - transactionDate) in whole days — the STOCK Act allows up to 45 days of lag, so "real-time" here means minutes-after-disclosure, not minutes-after-trade. DATE QUALITY: source dates are sometimes wrong. When a row's dates are impossible or implausible, disclosureLagDays is null and dateQuality names the reason: transaction_after_disclosure (transaction date later than disclosure date), future_transaction_date (transaction date in the future), or implausible_lag (lag above 1500 days, almost always a mistyped year). dateQuality is null on a normal row. Flagged rows are never dropped and their raw transactionDate and disclosureDate are returned unchanged. Genuine late filings with a long but plausible lag keep their disclosureLagDays. For per-politician or per-ticker rollups use GET /v1/congress/politicians, /v1/congress/politicians/{idOrSlug}, or /v1/congress/tickers/{ticker} (all Pro+). To check whether an insider cluster-buy lines up with a congressional purchase in the same ticker, use GET /v1/signals/convergence (Pro+); for the company's own profile use GET /v1/companies/{ticker} (Free). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching.
list_filingsRead-onlyList Form 4 filings with optional ticker, CIK and date filters. Returns a paginated list of Form 4 filings, newest filed first. Filter by ticker, cik, and a from/to filed-date window. Each entry carries the accession number, company ticker/name, period of report, filed date, acceptedAt (the precise UTC SEC-acceptance instant, null if not captured), documentUrl (the public SEC document URL), amendment type (Original/Amendment), and the count of non-superseded transactions in that filing. Use this for a company's filing HISTORY; use GET /v1/filings/recent for a live newest-first feed (it has no page parameter), and GET /v1/transactions when you want the individual trades rather than the filings that contain them. `limit` is accepted as an alias for `per_page`. Not plan-gated.
list_schedule13_dgRead-onlyList Schedule 13D/13G >5% beneficial-ownership crossings (Business plan+). Returns every parsed >5% beneficial-ownership crossing event for a company — filer/reporting person, ownership percent, event date, form type (13D or 13G), and amendment status — parsed from EDGAR's structured Schedule 13D/13G cover-page XML (mandatory since the SEC's 2023 rule, compliance date 2024-12-18). SC 13D indicates an "active" investor who may seek control or influence over the issuer and must file within five business days of crossing the threshold (shortened from 10 days by the SEC's 2023 rule amendments, effective 2024-02-05); SC 13G is a lighter, less frequent filing for "passive"/qualified investors (e.g. index funds), with its own accelerated deadlines effective 2024-09-30. A single accession can carry several rows when the filing is a joint filing (e.g. a fund, its general partner, and the individual who controls both) — each reporting person has its own ownership percent. Results exclude rows superseded by a later SC 13D/A or SC 13G/A amendment; the isAmendment field on a returned row indicates whether that row is itself an amendment, not whether it has since been superseded. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching.
list_webhooksRead-onlyList webhook subscriptions owned by the authenticated API key. Returns active webhook subscriptions under the authenticated API key, plus any the system auto-disabled after sustained delivery failures (disabledReason "delivery_failures") — those stay visible with disabledAt/disabledReason set so you can see why delivery stopped and reverse it with POST /v1/webhooks/{id}/enable. Subscriptions you deleted yourself (disabledReason "user_deleted") are never returned here again, and neither are legacy rows disabled before this field existed (disabledReason null) — both are permanently gone from this list. Each entry: subscription id, target URL, subscribed event types, creation date, active flag, disabledAt/disabledReason (both null while active), and the isReadOnly flag. Delivery health, all nullable: lastAttemptAt (when the most recent delivery attempt, first attempt or retry, finished; null if none yet), lastSuccessAt (when the most recent attempt that got a 2xx finished), lastStatusCode (HTTP status of the most recent attempt; null when no response was received, e.g. timeout or connection failure), and lastErrorClass (why the most recent attempt failed, null if it succeeded or none has been made: HTTP_4XX, HTTP_5XX, HTTP_OTHER, TIMEOUT, DNS, TLS, CONNECT_REFUSED, URL_REJECTED, CIRCUIT_OPEN or OTHER). These reflect only attempts made after this field was introduced; older subscriptions show null until their next attempt. authHeaderName is the name of the custom auth header sent with each delivery (null if none is set); its value is never returned. urlWarning is a string when the URL's host is a temporary tunnel or test host (for example trycloudflare.com, ngrok or webhook.site), otherwise null. Change a subscription's URL with PATCH /v1/webhooks/{id}. Does NOT return the signing secret again (it's shown once, at creation, by POST /v1/webhooks) — regenerate by deleting and recreating the subscription if it's lost. Requires a valid X-Api-Key (401 without one).
search_insidersRead-onlySearch insiders (officers, directors, 10% owners) by name. Searches insiders by name and returns a paginated list of matches with each insider's CIK, title, director/officer/10%-owner flags, and total filing count. Use this to resolve a person's name to their CIK before fetching their transaction history, career summary, or scorecard — the CIK returned here feeds directly into GET /v1/insiders/{cik}/transactions, /summary, and /scorecard. Omitting the name filter returns insiders in alphabetical order rather than performing a search. `limit` is accepted as an alias for `per_page`. Not plan-gated — available on the Free tier.
| get_insider_career_summary | Aggregate career rollup for one insider by CIK — first/last transaction dates, total shares/value bought vs sold, top companies traded, transaction-code breakdown, 10b5-1 plan split, and median post-trade returns (stored as fractions, e.g. 0.05 = +5%). Use this instead of get_insider_transactions when you want a pre-computed summary rather than raw rows; pair with the generated get_insider_scorecard for hit-rate statistics on their discretionary buys specifically. Requires Pro plan. | Read-only |
| check_usage | Snapshot of the authenticated API key's current usage — plan name, requests made today, daily limit, and all-time request count. Use this for a quick right-now check; use the generated get_usage_history tool for a daily trend over time, or get_key_activity for a per-request log. Free plan, works on every tier. | Read-only |
| verify_setup | Diagnostic check: confirms FORM4API_KEY is present and valid, the backend is reachable, and reports your current plan — returns a pass/fail/warn per check plus concrete next steps (where to get a key, how to upgrade) instead of a raw error. Run this first whenever another tool fails or returns a 401/402, to isolate whether the problem is configuration, plan, or a backend outage. No API key or parameters required. | Read-only |
| search | Resolve free-text input — a company name, ticker fragment, or a person's name — to a ticker or CIK. This is usually the right FIRST call when a user names a company or insider you don't already have an identifier for; feed the resulting ticker into get_company_overview/get_transactions or the CIK into get_insider_profile/get_insider_transactions. Returns two independently-ranked lists: `companies` (only ones with a resolved public ticker, ranked exact ticker match, then ticker-prefix, then name-prefix, then name-contains) and `insiders` (matched by splitting the query on whitespace so every token must match the name, e.g. "tim cook" matches SEC-style "Cook Timothy D"; ranked by total filing count, most active first). Each insider's `ticker` is a ticker associated with that insider (may be null). `q` must be 2-64 characters after trimming (400 QUERY_TOO_SHORT/QUERY_TOO_LONG). `limit` applies independently to each list, default 8, max 20. Free plan, no caching. | Read-only |
| research_company | Bundled insider-research context for one ticker in a single call — company profile, recent insider transactions (10b5-1 flagged), cluster buy/sell signals, monthly sentiment score, and a computed net buy/sell direction summary. Use this FIRST for any company insider-research question; it replaces 4 separate calls (get_company_overview, get_transactions, get_signals, get_sentiment) and degrades gracefully — if signals/sentiment require a plan you don't have, they come back null with an explanatory `_unavailable` note instead of failing the whole call. Company/transaction data works on Free; signals/sentiment sections require Business. Fetches all sections concurrently. Research and decision-support only, not investment advice. | Read-only |
| explain_signal | Explain why a signal fired: the insiders and trades counted, what was excluded, and the criteria (Business plan+). Use this when the question is WHY a named company's signal fired, or WHICH insiders and trades made up its cluster. It takes the ticker directly, so there is no need to call GET /v1/signals first; that endpoint is for finding which companies have signals. Reconstructs the full evidence behind one company's insider signal from GET /v1/signals: the detection criteria (5-day cluster window, 3-insider threshold, 90-day ratio window, 10b5-1 and derivative-row exclusion), the list of cluster buyers and sellers (each with their role and individual trades in the window), trades that were excluded from the cluster count and why (10b5-1 plan, derivative security, or superseded by amendment), and the raw buy/sell share totals behind the 90-day ratio. Use this to audit or debug a specific signal rather than to scan many companies (use GET /v1/signals for that). This is a LIVE reconstruction from current non-superseded data, computed on every request (no caching) — it can differ slightly from the originally stored signal if trades were amended afterward. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). Returns 404 COMPANY_NOT_FOUND if the ticker isn't tracked, 404 SIGNAL_NOT_FOUND if no signal exists for the given/most-recent date, or 400 INVALID_DATE if `date` isn't YYYY-MM-DD. | Read-only |
| get_company_peers | Peer companies in the same SIC industry, ranked by recent Form 4 activity. Returns other tracked companies in the same industry as the given ticker, for "similar companies" links. Candidates share the company's 4-digit SIC code (matchLevel "sic4"); when too few of those have Form 4 filings in the last 12 months to fill `limit`, companies sharing only the 2-digit SIC prefix (matchLevel "sic2") are added. Active peers (at least one Form 4 filing dated in the trailing 12 months, `filings12m`) come first, sic4 before sic2, each tier ordered by `filings12m` descending then ticker ascending; peers with no filings in the window are appended last only if there is still room. The company itself is excluded, as are companies without a ticker or without a usable SIC code, and any other company holding the same ticker. When several companies hold the ticker, the one GET /v1/companies/{ticker} returns is used. `basis` is "sic4", "sic2" or "mixed" according to the match levels present in `peers`, and "none" when `peers` is empty (including when the company has no SIC code, which is 200 with an empty list, not 404). `limit` defaults to 10 and is clamped to 1-25. Returns 404 NOT_FOUND if the ticker isn't tracked. Not plan-gated. Results are cached for a few hours. | Read-only |
| get_company_returns_summary | What happened after insiders traded at a company (Pro plan+). Summarises price returns after insider open-market trades at one company, for buys and for discretionary sells, next to a site-wide baseline computed the same way. The unit is an event: one per company per filing date and side. An event's return at a horizon (1m, 3m, 6m = 21, 63, 126 trading days) is the average of that event's rows, measured from the filing-date close. Included: open-market purchases (P) and sales (S) that are not superseded, not derivative and carry no value-quality flag, filed since 2023-05-22; sells exclude 10b5-1 plan sales. An event counts toward a horizon only once that horizon has matured. `nEvents` is the number of events with a 3m return; `n1m` and `n6m` are the counts for the other horizons. Medians are null when that horizon has fewer than 10 events; `mean3m` and `hitRate3m` (share of events with a positive 3m return) are null when `nEvents` is below 10. `events` lists the individual events, newest first, only when `nEvents` is between 3 and 9, otherwise it is empty. `suppressed` is true when the median filing-date close across the company's buy and sell events is below $1, and `lowPrice` is true when that median is from $1 up to below $5; the price itself is not returned. Market comparison: each event's stock return is compared with the S&P 500 (SPY) return over the same window, from the stock's filing-date close bar to the stock's own bar at the horizon, matched by exact date; the event's excess return is the stock return minus the SPY return, computed only when both exist (events without an SPY return are left out, never filled in). `nExcess1m`, `nExcess3m` and `nExcess6m` count the events with an excess return per horizon; `excessMedian1m`, `excessMedian3m` and `excessMedian6m` are null when that count is below 10; `excessMean3m` and `beatMarket3m` (share of events whose 3m stock return exceeded SPY's) are null when `nExcess3m` is below 10. Each item in `events` carries `excess3m`. Not adjusted for company size or sector. The `baseline` block (nEvents, median3m, mean3m, hitRate3m, plus excessMedian3m, excessMean3m and beatMarket3m for buys and sells) is site-wide. All return fields are FRACTIONS, 0.05 means +5%. Requires Pro plan or higher (402 PLAN_REQUIRED on Free). Returns 404 NOT_FOUND if the ticker isn't tracked; when several companies hold the ticker, the one GET /v1/companies/{ticker} returns is used. Cached for 24 hours per ticker, and the baseline is cached for 24 hours, so `asOf` can lag the latest filings by up to a day or two. | Read-only |
| get_congress_politician | Get one politician's congressional trading profile (Pro plan+). Returns one politician's profile — identity fields, total/buy/sell trade counts, most recent trade's disclosure date, their top N most-traded tickers (by trade count), and their N most recent trades (same shape as GET /v1/congress/trades). Use this for a one-call politician overview rather than paging /v1/congress/trades?politician= yourself. Accepts either a bioguide ID (e.g. "P000197") or the politician's URL slug (e.g. "nancy-pelosi") in the path, matched case-insensitively against whichever field applies. Returns 404 NOT_FOUND if neither matches. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Query runs live — no caching. | Read-only |
| get_congress_ticker_rollup | Which politicians traded a ticker, with net buy/sell counts (Pro plan+). Returns every politician who has a non-superseded congressional trade in the given ticker, each with their trade/buy/sell counts, plus ticker-level totals. Optional window_days restricts to trades with a transactionDate in the trailing N days; omit for all-time. A ticker with no congress trades returns 200 with an empty politicians array and zero counts rather than 404 — there is no separate ticker/company entity in this dataset to 404 against. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Query runs live — no caching. | Read-only |
| get_convergence_signals | Insider cluster-buy x congressional-purchase convergence (Pro plan+). Returns the tickers where an insider cluster-buy (InsiderSignal.IsClusterBuy) and at least one non-superseded congressional PURCHASE happened within window_days of EACH OTHER, restricted to convergences where the MORE RECENT of the pair's two dates is within a trailing lookback_days (so this surfaces CURRENT convergences, not ancient history). DEFINITION: for each result, insider.signalDate is the SignalDate of the qualifying cluster-buy signal with the most recent date (insider.insiderCount is that same signal's count — never summed or maxed across multiple signals), and congress is every non-superseded congressional purchase that paired with at least one qualifying cluster-buy (not every purchase in the window — only the ones that actually paired). firstSeen/lastSeen are the earliest/most recent dates among all qualifying insider and congress dates for that ticker. STRENGTH is documented arithmetic, NOT a black-box or predictive/ML score: strength = (distinct congressional purchasers among the qualifying legs) x (the representative signal's insiderCount) — a plain multiplication of two observed counts, nothing more. HONESTY: every congress leg always carries both amountLow and amountHigh (STOCK Act discloses ranges, never exact figures — never combined into a fabricated midpoint) and disclosureLagDays = (disclosureDate - transactionDate). Congressional trades whose dates are impossible or implausible (a dateQuality flag: transaction_after_disclosure, future_transaction_date or implausible_lag) are excluded from detection, so they never appear as legs and never create a convergence; dateQuality is therefore null on every leg returned here and disclosureLagDays is always set. Congressional trades are disclosed up to 45 days after the actual trade under the STOCK Act, so this endpoint is detection/monitoring of what insiders AND members of Congress have DISCLOSED buying, not a claim of predictive edge, alpha, or win rate — no performance numbers are computed or implied anywhere in this response. window_days and lookback_days are both caller-overridable with clamps (see each parameter's own description for the exact bounds). Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching. | Read-only |
| get_data_quality | Public data-quality, freshness and coverage metrics for the whole dataset. Returns public, keyless metrics on data freshness, ingestion latency, corpus coverage, and post-trade returns coverage — use this to check whether the dataset is current before relying on it (e.g. confirm Form 4 ingestion isn't stalled, or that price data isn't stale), not to look up any single company, insider, or transaction. Includes: most recent Form 4 processed timestamp and median/p95 filing-accepted-to-processed latency in seconds, latest price-bar date and how many days behind it is, total companies/transactions tracked plus filing counts by form type (4, 144, 13F-HR), the percentage of 13F CUSIPs resolved to a ticker, and the percentage of eligible transactions with fully computed post-trade returns. Takes no parameters. Cached for 30 minutes; no API key or plan required. | Read-only |
| get_insider_directory | Browse insiders alphabetically by surname. Returns the A-Z rail with a count per letter, plus one page of insiders under the requested letter. Omit `letter` to get the rail and totals with no rows. Names come from EDGAR surname-first ("HENNEMAN JOHN B III"), so alphabetical order is order by surname. Casing in the source is inconsistent and is not normalised here. This lists only insiders with at least 3 non-superseded transactions, capped at the 5,000 most active — the same set as the insiders sitemap shard, so the two cannot drift. To find someone outside that set, use GET /v1/insiders?name= which searches every filer. Rebuilt daily; `refreshedAt` reports when. `limit` is accepted as an alias for `per_page`. Not plan-gated. One row per FILER GROUP. A fund group files a single Form 4 listing several reporting owners — the fund, its GP, its management company — and each is a real EDGAR filer with its own CIK. Listing all of them spent about 11% of this capped surface describing the same actors more than once, so browse shows one per group and `filerGroupSize` says how many others share those exact transactions. The others are not hidden: each keeps its own profile and is still returned by GET /v1/insiders?name=. | Read-only |
| get_insider_leaderboard | Ranked leaderboard of insiders by buy track-record (Business plan+). Returns the top insiders ranked by historical buy performance — same scored-buy methodology as GET /v1/insiders/{cik}/scorecard, applied across the whole corpus rather than one insider. Use this to discover which insiders have the best track record; use the per-insider scorecard once you have a specific CIK. Scores use absolute return (NOT market-adjusted) — a hit is a scored buy with a positive 3m (or 6m) return anchored at the filing-date close. Only discretionary open-market buys (P-code, not 10b5-1, not derivative) with a matured return are counted. Insiders with fewer than min_trades (floor 5) scored buys are excluded. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). Results are cached for 1 hour per unique parameter combination. | Read-only |
| get_insider_scorecard | Get insider buy track-record scorecard (Pro plan+). Returns the historical hit rate and average/median return of an insider's discretionary open-market buys (TransactionCode=P, excluding 10b5-1 plans and derivatives), plus their best and worst scored buy. Scores use absolute return (NOT market-adjusted) anchored at the filing-date close. A 'hit' is a scored buy whose 3m (or 6m) return is positive. Use this over GET /v1/insiders/{cik}/summary when you specifically want a scored track record (with a sample-sufficiency guard) rather than raw totals; use GET /v1/insiders/leaderboard (Business+) to rank many insiders by this same methodology. Score fields (hitRate3m, avgReturn3m, etc.) are null when the insider has fewer than 5 matured scored buys (sampleSufficient=false), preventing misleading statistics from small samples. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). Returns 404 NOT_FOUND if the CIK isn't tracked. Computed live — no caching. Note: all return fields (HitRate3m, AvgReturn3m, MedianReturn3m, etc.) are stored as FRACTIONS — 0.05 means +5%, -0.10 means -10%. | Read-only |
| get_key_activity | Recent, per-request API activity log for the authenticated key. Returns the most recent HTTP requests made with the authenticated API key, most recent first, including the endpoint path, response status code, duration in milliseconds, and timestamp. Use this to debug integration issues — confirm a specific call reached the API, check for repeated 4xx/5xx responses, or spot slow requests — rather than for usage trends; for aggregate daily counts use GET /v1/keys/usage/history instead. Requires a valid X-Api-Key (401 without one). Limited to 30 requests per minute per API key, shared with the other /v1/keys/usage routes (429 KEY_USAGE_RATE_LIMIT). | Read-only |
| get_public_stats | Public corpus-wide statistics — no API key required. Returns headline dataset totals: filing count, transaction count, tracked companies, institutional holdings rows, Form 144 and Form 13F-HR filing counts, the earliest filing date in the corpus, the most recent quarter's total 13F-HR reported AUM in USD, and measured ingestion latency (median/p95 seconds from SEC acceptance to our processing, trailing 7 days). Use this for corpus-wide totals (e.g. a marketing/status widget), not for per-company or per-insider data — those live under GET /v1/companies and GET /v1/insiders. For freshness and coverage-quality metrics (is ingestion stalled, is price data stale) use GET /v1/data-quality instead. Takes no parameters. No API key or plan required. Cached for ~12 hours. | Read-only |
| get_status_history | Measured uptime history for the public status page — trailing 90-day daily breakdown. Returns a daily breakdown of measured API uptime over a trailing 90-day window, computed from an internal heartbeat probe that runs every 5 minutes and performs the same DB-connectivity check as GET /health/ready. Each day in the `days` array reports the number of 5-minute slots expected to have elapsed (288 for a complete past day, pro-rated for the feature's first day and for today's partial day), how many of those slots recorded a healthy heartbeat, and the resulting uptime percentage for that day — plus an overall percentage (`overallPct`) across the whole window. `start` is the earliest date included: either the date of the very first heartbeat ever recorded, or 89 days before today once more than 90 days of history exist. Days before that are never returned. Use this to render an uptime history / status bar; for live corpus freshness use GET /v1/data-quality instead. Takes no parameters. Cached for ~5 minutes; no API key or plan required. | Read-only |
| get_usage_history | Daily request counts for the authenticated key over a trailing window. Returns a daily time series of request counts for the authenticated API key over the trailing N days — one data point per calendar day (UTC). Use this to plot usage trends or check rate-limit headroom over time. For a single current-day snapshot (today's count, plan limit, reset time) use GET /v1/keys/usage instead; for a raw request-by-request log use GET /v1/keys/usage/activity. Requires a valid X-Api-Key (401 without one). Limited to 30 requests per minute per API key, shared with the other /v1/keys/usage routes (429 KEY_USAGE_RATE_LIMIT). | Read-only |
| get_webhook_events | Replay webhook delivery events since a given timestamp (default: last 24h). Returns up to 500 delivery attempts across all of the authenticated key's subscriptions since `since` (default: last 24 hours), most recent first — delivery id, subscription id, event type, attempt count, delivered-at/next-retry-at timestamps, last HTTP status code from the receiving endpoint, whether the delivery is dead (exhausted all retries), and the event payload. Use this to reconcile missed webhook deliveries (e.g. after an outage on your receiving endpoint) rather than relying solely on push delivery. `since` cannot be more than 30 days in the past. Requires a valid X-Api-Key (401 without one). `eventType` on each row is the exact name accepted by POST /v1/webhooks and sent in the X-Event-Type delivery header (e.g. "CongressTradeFiled", "ConvergenceSignal" — not dotted names). `payload` is null and `payloadRedacted` is true for any event type above your current plan (CongressTradeFiled requires Starter, ConvergenceSignal requires Pro) — delivery history outlives the plan that created it, so payloads are checked against the plan you are on now, not the plan you had when you subscribed. | Read-only |
| health_ingestion | Ingestion-death detector — Form 4 freshness, parse-queue health, price-feed freshness. Returns live (uncached) ingestion health: Form 4 processing freshness (stale after 90 minutes during the weekday 06:00-22:00 US Eastern EDGAR activity window, counted from the later of the last filing and the 06:00 opening; outside it, 14 hours or the time since the window last closed plus 90 minutes, whichever is longer, so a normal weekend never reads as stale), parse-queue health (stale when >50 jobs are pending AND the oldest has waited >45 minutes), and daily price-feed freshness (degraded, never stale, when price bars are >5 days behind). Overall status is "healthy", "degraded" (price feed only), or "stale" (Form 4 or queue). Returns HTTP 503 when stale, 200 otherwise — safe to point an external uptime monitor at directly. No API key or plan required. | Read-only |
| list_companies | List companies with a public ticker, sorted by name or total filings. Returns a single page of companies that have a tracked public ticker — for browsing or building a company picker, not for searching by name or CIK (there is no full-text search here; use GET /v1/companies/{ticker} to fetch one company by its exact ticker). Each entry includes the company's CIK, name, ticker, exchange, total filing count, and distinct insider count. There is no page parameter — this endpoint always returns the top `limit` companies by the chosen sort order. Not plan-gated. | Read-only |
| list_congress_politicians | Ranked rollup of politicians by congressional trade activity (Pro plan+). Returns a paginated list of politicians who have at least one non-superseded congressional trade, each with total/buy/sell counts (sells include both Sale and PartialSale; Exchange trades count only toward total) and their most recent trade's disclosure date. Ordered by total trade count descending, ties broken by most recently disclosed. Use this to discover active traders; for one politician's full profile (including their most-traded tickers and recent trades) use GET /v1/congress/politicians/{idOrSlug}. Requires Pro plan or higher (402 PLAN_REQUIRED on Free/Starter). `limit` is accepted as an alias for `per_page`. Query runs live — no caching. | Read-only |
| list_congress_trades | Query congressional STOCK Act trades (Free+, plan-clamped disclosure window). Returns a paginated JSON list of congressional periodic-transaction-report trades, most recently DISCLOSED first, with non-superseded rows only (amended-away rows never appear). COVERAGE — HOUSE ONLY TODAY: every trade in this dataset comes from the U.S. House Clerk's PTR index. Senate eFD (efdsearch.senate.gov) returns 403 to datacenter traffic, so no Senate filings are ingested yet. chamber=Senate remains a valid filter but matches nothing and returns the response header X-Coverage-Note: chamber-not-covered, so an empty result is never ambiguous. Scanning by chamber should treat that header as "not covered", not as "no trades". PLAN-CLAMPED WINDOW: this endpoint is open to every plan, but how far back you can see is clamped on disclosureDate — Free sees only trades disclosed in the last 30 days, Starter the last 366 days, Pro/Business/Enterprise unlimited history. Passing an older disclosure_date_from than your plan allows does not extend the window — the floor always wins. Filters: ticker, politician (bioguideId, exact), party (free-text, case-insensitive exact match — not a fixed enum), chamber (House|Senate — see the coverage note above), state (2-letter code), transaction_type (purchase|sale|partial_sale|exchange), min_amount (range-aware — matches AmountLow >= value, never a fabricated midpoint), transaction_date_from/to, disclosure_date_from/to. Every row always carries BOTH amountLow and amountHigh (STOCK Act discloses ranges, never exact figures) and disclosureLagDays = (disclosureDate - transactionDate) in whole days — the STOCK Act allows up to 45 days of lag, so "real-time" here means minutes-after-disclosure, not minutes-after-trade. DATE QUALITY: source dates are sometimes wrong. When a row's dates are impossible or implausible, disclosureLagDays is null and dateQuality names the reason: transaction_after_disclosure (transaction date later than disclosure date), future_transaction_date (transaction date in the future), or implausible_lag (lag above 1500 days, almost always a mistyped year). dateQuality is null on a normal row. Flagged rows are never dropped and their raw transactionDate and disclosureDate are returned unchanged. Genuine late filings with a long but plausible lag keep their disclosureLagDays. For per-politician or per-ticker rollups use GET /v1/congress/politicians, /v1/congress/politicians/{idOrSlug}, or /v1/congress/tickers/{ticker} (all Pro+). To check whether an insider cluster-buy lines up with a congressional purchase in the same ticker, use GET /v1/signals/convergence (Pro+); for the company's own profile use GET /v1/companies/{ticker} (Free). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching. | Read-only |
| list_filings | List Form 4 filings with optional ticker, CIK and date filters. Returns a paginated list of Form 4 filings, newest filed first. Filter by ticker, cik, and a from/to filed-date window. Each entry carries the accession number, company ticker/name, period of report, filed date, acceptedAt (the precise UTC SEC-acceptance instant, null if not captured), documentUrl (the public SEC document URL), amendment type (Original/Amendment), and the count of non-superseded transactions in that filing. Use this for a company's filing HISTORY; use GET /v1/filings/recent for a live newest-first feed (it has no page parameter), and GET /v1/transactions when you want the individual trades rather than the filings that contain them. `limit` is accepted as an alias for `per_page`. Not plan-gated. | Read-only |
| list_schedule13_dg | List Schedule 13D/13G >5% beneficial-ownership crossings (Business plan+). Returns every parsed >5% beneficial-ownership crossing event for a company — filer/reporting person, ownership percent, event date, form type (13D or 13G), and amendment status — parsed from EDGAR's structured Schedule 13D/13G cover-page XML (mandatory since the SEC's 2023 rule, compliance date 2024-12-18). SC 13D indicates an "active" investor who may seek control or influence over the issuer and must file within five business days of crossing the threshold (shortened from 10 days by the SEC's 2023 rule amendments, effective 2024-02-05); SC 13G is a lighter, less frequent filing for "passive"/qualified investors (e.g. index funds), with its own accelerated deadlines effective 2024-09-30. A single accession can carry several rows when the filing is a joint filing (e.g. a fund, its general partner, and the individual who controls both) — each reporting person has its own ownership percent. Results exclude rows superseded by a later SC 13D/A or SC 13G/A amendment; the isAmendment field on a returned row indicates whether that row is itself an amendment, not whether it has since been superseded. Requires Business plan or higher (402 PLAN_REQUIRED on Free/Starter/Pro). `limit` is accepted as an alias for `per_page`. Query runs live against the database — no caching. | Read-only |
| list_webhooks | List webhook subscriptions owned by the authenticated API key. Returns active webhook subscriptions under the authenticated API key, plus any the system auto-disabled after sustained delivery failures (disabledReason "delivery_failures") — those stay visible with disabledAt/disabledReason set so you can see why delivery stopped and reverse it with POST /v1/webhooks/{id}/enable. Subscriptions you deleted yourself (disabledReason "user_deleted") are never returned here again, and neither are legacy rows disabled before this field existed (disabledReason null) — both are permanently gone from this list. Each entry: subscription id, target URL, subscribed event types, creation date, active flag, disabledAt/disabledReason (both null while active), and the isReadOnly flag. Delivery health, all nullable: lastAttemptAt (when the most recent delivery attempt, first attempt or retry, finished; null if none yet), lastSuccessAt (when the most recent attempt that got a 2xx finished), lastStatusCode (HTTP status of the most recent attempt; null when no response was received, e.g. timeout or connection failure), and lastErrorClass (why the most recent attempt failed, null if it succeeded or none has been made: HTTP_4XX, HTTP_5XX, HTTP_OTHER, TIMEOUT, DNS, TLS, CONNECT_REFUSED, URL_REJECTED, CIRCUIT_OPEN or OTHER). These reflect only attempts made after this field was introduced; older subscriptions show null until their next attempt. authHeaderName is the name of the custom auth header sent with each delivery (null if none is set); its value is never returned. urlWarning is a string when the URL's host is a temporary tunnel or test host (for example trycloudflare.com, ngrok or webhook.site), otherwise null. Change a subscription's URL with PATCH /v1/webhooks/{id}. Does NOT return the signing secret again (it's shown once, at creation, by POST /v1/webhooks) — regenerate by deleting and recreating the subscription if it's lost. Requires a valid X-Api-Key (401 without one). | Read-only |
| search_insiders | Search insiders (officers, directors, 10% owners) by name. Searches insiders by name and returns a paginated list of matches with each insider's CIK, title, director/officer/10%-owner flags, and total filing count. Use this to resolve a person's name to their CIK before fetching their transaction history, career summary, or scorecard — the CIK returned here feeds directly into GET /v1/insiders/{cik}/transactions, /summary, and /scorecard. Omitting the name filter returns insiders in alphabetical order rather than performing a search. `limit` is accepted as an alias for `per_page`. Not plan-gated — available on the Free tier. | Read-only |
Configuration
FORM4API_KEYRequiredSecretForm4API key — get a free one (500 req/day, no card) at https://www.form4api.com
| Name | Description | |
|---|---|---|
| FORM4API_KEY | Form4API key — get a free one (500 req/day, no card) at https://www.form4api.com | RequiredSecret |
Automated checks
Ran Nov 12, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packagePass
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledPass
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesPass
README
View on GitHubform4api-mcp
Production-grade SEC Form 4 insider trading data for any MCP-compatible AI assistant — amendment-aware, 10b5-1 clean, with Form 144 + institutional 13F-HR overlay, plus congressional STOCK Act trades and insider/Congress convergence — 39 tools + 6 ready-made research prompts
An MCP server that exposes the hosted Form4API REST API to Claude Desktop, Cursor, Windsurf, VS Code, Codex CLI, and any other MCP-compatible client. Configured once, your LLM can answer questions about insider trading, institutional positioning, and intent-to-sell filings directly during research sessions.
Four data-quality claims no scraping-based alternative can make:
- 🛡 Amendment-aware — Form 4/A amendments are reconciled automatically. No double-counting when an insider corrects a filing.
- 🎯 10b5-1 clean — every transaction flagged as pre-scheduled (10b5-1 plan) or discretionary. Cluster signals exclude planned trades by construction.
- 📜 Form 144 intent-to-sell — 125K+ Form 144 filings indexed. Catch insider sales ~2 days before they hit Form 4.
- 🏛 Institutional × insider join — every transaction carries the current 13F-HR ownership context (top-3 holders, AUM trend) in the same response: no second call, no client-side join. Among the self-serve SEC data APIs we've surveyed, none return both sides in one query — sec-api.io and Kaleidoscope both ship 13F and insider data as separate endpoints.
Quick install
1. Get a free API key
Go to www.form4api.com → Sign in → Dashboard. Free plan includes 500 requests/day, no credit card required.
2. Add to your MCP client
Claude Desktop — edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"form4api": {
"command": "npx",
"args": ["-y", "form4api-mcp"],
"env": {
"FORM4API_KEY": "YOUR_API_KEY"
}
}
}
}
Restart the client. The tools appear automatically.
Claude Code (CLI):
claude mcp add form4api -- npx -y form4api-mcp
…then set FORM4API_KEY in your shell or in ~/.claude/mcp.json.
Cursor — edit ~/.cursor/mcp.json (user-level) or .cursor/mcp.json (workspace-level):
{
"mcpServers": {
"form4api": {
"command": "npx",
"args": ["-y", "form4api-mcp"],
"env": {
"FORM4API_KEY": "YOUR_API_KEY"
}
}
}
}
Restart Cursor. The tools appear automatically.
Windsurf — edit ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"form4api": {
"command": "npx",
"args": ["-y", "form4api-mcp"],
"env": {
"FORM4API_KEY": "YOUR_API_KEY"
}
}
}
}
Restart Windsurf. The tools appear automatically.
VS Code — edit .vscode/mcp.json (workspace-level). Note: VS Code uses the servers key (not mcpServers):
{
"servers": {
"form4api": {
"command": "npx",
"args": ["-y", "form4api-mcp"],
"env": {
"FORM4API_KEY": "YOUR_API_KEY"
}
}
}
}
Restart VS Code. The tools appear automatically.
Codex CLI — config is TOML at ~/.codex/config.toml:
[mcp_servers.form4api]
command = "npx"
args = ["-y", "form4api-mcp"]
env = { FORM4API_KEY = "YOUR_API_KEY" }
Verify it works
Ask your LLM to run the verify_setup tool — it confirms your API key is valid and the MCP server is reachable, or returns the exact fix steps.
Example: "Run the verify_setup tool to confirm the MCP is configured correctly."
Try before you commit a key
get_public_stats is a keyless tool — it works with no FORM4API_KEY set. Try it first to preview live data coverage before signing up:
FORM4API_KEY="" npx form4api-mcp
Once you like what you see, sign up for a free key at www.form4api.com → set FORM4API_KEY → all tools unlock.
3. Or run directly
FORM4API_KEY=YOUR_API_KEY npx form4api-mcp
Available tools (39)
Search
| Tool | Description | Plan |
|---|---|---|
search |
Resolve a company name, ticker fragment, or a person's name to a ticker or CIK — the usual first call when you don't already have an identifier. Returns separately-ranked companies and insiders lists |
Free |
Form 4 insider trading
| Tool | Description | Plan |
|---|---|---|
research_company |
Bundled insider-research context for one ticker in a single call — company profile, recent transactions, cluster signals, sentiment, and a computed buy/sell direction summary. Replaces 4 separate calls and degrades gracefully when a section needs a higher plan | Free (signals/sentiment sections need Business) |
get_transactions |
Search insider transactions — filter by ticker, insider, date range, transaction codes or whole categories (exclude_category=derivatives), 10b5-1 plan trades, a dollar floor (min_value), the insider's role (role: officer, director or ten_percent_owner), ordering by trade size (sort=value; default is newest first), the 13F ownership trend (inst_ownership_trend), or use significant=true for real discretionary buys/sells only. Pro adds the remaining trade-size screens (max_value, min_shares, max_shares) and post-trade-return screening (min_return_1d…max_return_6m, has_returns; returns are fractions, 0.05 = +5%). Paging depth is plan-limited — see Plans |
Free |
get_recent_filings |
Most recent Form 4 filings, optionally filtered by ticker | Free |
list_filings |
Form 4 filings as a paginated list, newest filed first — filter by ticker, cik, or a filed-date window. Use this to page through filings; get_recent_filings is the unfiltered head of the same feed |
Free |
get_filing |
Single filing by accession number | Free |
get_insider_profile |
Insider profile — name, title, director/officer/10pct owner flags | Free |
get_insider_transactions |
All transactions for a specific insider (by CIK) | Free |
get_company_overview |
Company profile — name, CIK, SIC sector, state, website, filing counts | Free |
get_company_insiders |
All insiders who have filed Form 4s for a company | Free |
get_company_peers |
Similar companies in the same SIC industry (4-digit first, topped up from the 2-digit prefix), ranked by Form 4 filings in the last 12 months | Free |
list_companies |
List companies, sorted by name or filing count | Free |
get_insider_directory |
Browse insiders alphabetically by surname — an A-Z rail with a count per letter, plus one page of insiders under the requested letter | Free |
get_insider_career_summary |
Aggregate career rollup: total bought/sold, top companies, 10b5-1 split, return averages | Pro |
get_insider_scorecard |
Buy track-record scorecard for an insider (CIK) — hit rate and avg/median return on discretionary open-market buys; null when fewer than 5 matured samples | Pro |
get_insider_leaderboard |
Top insiders ranked by hit_rate or avg_return; filter by horizon (3m/6m), min_trades, and limit |
Business |
Signals + sentiment
| Tool | Description | Plan |
|---|---|---|
get_signals |
Cluster buy/sell signals — multiple insiders at the same company in the same direction. Excludes 10b5-1 trades automatically | Business |
get_sentiment |
MSPR-style monthly sentiment score per ticker (-100 to +100). 10b5-1 excluded so the score reflects real insider conviction | Business |
Form 144 + institutional
| Tool | Description | Plan |
|---|---|---|
get_form144 |
Notice-of-proposed-sale filings — early signal ~2 days before Form 4 sale lands | Business |
get_holdings |
Institutional positions from Form 13F-HR (filter by ticker, CUSIP, manager, quarter, min value) | Business |
get_managers |
Institutional manager index with latest AUM | Business |
explain_signal |
Explain why a signal fired — the insiders and trades counted, exclusions, and criteria | Business |
get_data_quality |
Public data-quality, freshness and coverage metrics | Free |
list_schedule13_dg |
Schedule 13D/13G >5% beneficial-ownership crossings for a company — filer, ownership percent, event date, form type | Business |
Congress + convergence
| Tool | Description | Plan |
|---|---|---|
list_congress_trades |
Congressional STOCK Act trades (periodic transaction reports) — filter by ticker, politician, party, chamber, state, transaction type, min amount, or date range. Coverage is U.S. House only — Senate eFD blocks datacenter traffic, so chamber=Senate matches nothing and the response carries X-Coverage-Note: chamber-not-covered. Every row carries amountLow/amountHigh (disclosed ranges, never a fabricated midpoint) and disclosureLagDays — up to 45 days under the STOCK Act, so "real-time" here means minutes-after-disclosure, not minutes-after-trade |
Free (30-day disclosure window; Starter 366 days; Pro+ unlimited history) |
list_congress_politicians |
Ranked rollup of politicians by congressional trade activity — total/buy/sell counts, most recent disclosure | Pro |
get_congress_politician |
One politician's full profile by bioguide ID — totals, top traded tickers, most recent trades | Pro |
get_congress_ticker_rollup |
Which politicians traded a given ticker, with net buy/sell counts | Pro |
get_convergence_signals |
Tickers where an insider cluster-buy and a congressional purchase happened within a trailing window of each other. strength is documented arithmetic (distinct congressional purchasers × the signal's insider count) — never a black-box or predictive score. No performance/alpha claims are computed or implied |
Pro |
Utility
| Tool | Description | Plan |
|---|---|---|
check_usage |
Your API key usage stats and current plan | Free |
get_key_activity |
Recent API requests for this key | Free |
get_usage_history |
Daily request counts for the last N days | Free |
search_insiders |
Substring search on insider names | Free |
list_webhooks |
List your webhook subscriptions, with delivery health per webhook (lastAttemptAt, lastSuccessAt, lastStatusCode, lastErrorClass) and auto-disable state. Read-only |
Free |
get_webhook_events |
Replay webhook delivery events since a timestamp. Read-only | Free |
verify_setup |
Verify the MCP is configured correctly — confirms API key is valid and server is reachable | Free |
get_public_stats |
Public corpus-wide totals (filings, transactions, companies, 13F-HR AUM, ingestion latency) — no API key required | Free (keyless) |
get_status_history |
Trailing 90-day daily uptime history for the public status page | Free (keyless) |
health_ingestion |
Live ingestion-health check — Form 4 freshness, parse-queue backlog, price-feed staleness | Free (keyless) |
Prompts (6)
Beyond the 39 tools, this MCP ships 6 prompts — ready-made research recipes that a client can list (prompts/list) and load (prompts/get) so you don't have to hand-assemble the right tool sequence yourself. Each one tells the LLM exactly which SEC Form 4 / Form 144 / 13F-HR tools to call, in what order, and how to read plan-gated results.
| Prompt | Args | What it does |
|---|---|---|
insider_monitor |
ticker |
Recent SEC Form 4 insider activity for a ticker — transactions (10b5-1 flagged), cluster signals, sentiment — summarized as buy/sell conviction with post-trade-return context |
cluster_buy_scan |
days (default 7) |
Market-wide scan of recent cluster-buy signals, 10b5-1 excluded, ranked by conviction (insider count + $ value), each with a sentiment score |
form144_early_warning |
ticker (optional) |
Pending Form 144 notice-of-proposed-sale filings cross-referenced against recent Form 4 sells — flags discretionary (non-10b5-1) notices as the highest-signal early warnings, ~2 days ahead of the sale |
exec_conviction_check |
insider (name or CIK) |
An insider's career track record — total bought/sold, historical post-trade returns on discretionary buys, and whether their buying has historically beaten their scheduled 10b5-1 selling |
institutional_insider_overlap |
ticker |
Combines 13F-HR institutional holders with recent insider transactions to spot where smart money and insiders agree or diverge |
post_selloff_buys |
min_return (default 0.05) |
Screens insider buys with post-trade-return filters to surface historically-successful dip-buying patterns |
These map to the recipe workflows scraping-based Form 4 MCPs don't offer — each one leans on data this MCP alone exposes (10b5-1 flags, Form 144, 13F-HR join, per-insider return scoring). Plan requirements are honored honestly: prompts that touch Business-plan tools (get_signals, get_sentiment, get_form144, get_holdings, get_managers) or Pro-plan tools (get_insider_career_summary, get_insider_scorecard) instruct the LLM to surface the structured upgrade_required response rather than silently failing.
In Claude Desktop, prompts surface as a / slash-command picker; in Claude Code or other MCP clients, ask the assistant to "use the insider_monitor prompt for NVDA" (or similar) and it will fetch and follow the recipe.
Example prompts
Configured? Ask your LLM:
Quality-led (these require our amendment-aware, 10b5-1 clean, joined dataset):
- "Show me cluster buy signals from this week — discretionary only, no planned trades"
- "Which companies have insiders buying while institutional ownership is increasing this quarter?"
- "Are there any Form 144 filings at NVDA suggesting upcoming sales?"
- "What's the monthly insider sentiment for TSLA over the last 6 months, with 10b5-1 plans excluded?"
- "Berkshire Hathaway's top 13F-HR holdings — what did they add or trim this quarter?"
Standard insider research:
- "What insider trades happened at NVDA in the last 30 days, excluding 10b5-1 plans?"
- "What is Tim Cook's career insider-trading summary?"
- "Show me all open-market purchases over $1M at Tesla in 2026"
- "What has the CFO of Microsoft been doing with their shares this year?"
Why this MCP vs scraping-based alternatives
Some MCPs in this space scrape free public sites (e.g. openinsider.com) for Form 4 data. That's fine for a quick prototype but the data layer they give your LLM has structural gaps:
| form4api-mcp | Scraping-based MCPs | |
|---|---|---|
| Form 4/A amendment handling | ✅ reconciled automatically | ❌ double-counts |
| 10b5-1 plan flag | ✅ exposed on every transaction | ❌ planned + discretionary mixed |
| Form 144 intent-to-sell | ✅ 125K+ filings | ❌ not exposed |
| Institutional × insider join | ✅ top-3 holders + AUM trend per transaction | ❌ insider only |
| Sentiment (10b5-1 excluded) | ✅ MSPR-style score | ❌ planned trades pollute score |
| Source resilience | ✅ hosted API contract | ❌ breaks when source HTML changes |
| Webhooks / production delivery | ✅ HMAC + retry + DLQ | ❌ MCP-only, no fallback |
| SDKs | ✅ Python + JS | ❌ MCP-only |
| Commercial support | ✅ | ❌ |
If your LLM session is the start of a real research workflow that eventually wants production delivery, picking the MCP that has a graduation path matters.
Beyond MCP — when you need more
Webhook tools are read-only. The MCP can list your webhooks (list_webhooks) and replay delivery events (get_webhook_events), nothing more. Creating, deleting, enabling and testing webhooks, and setting a custom auth header, are done through the REST API or the dashboard. This is deliberate: the MCP exposes no write tools for webhooks.
The MCP is the easiest entry point. When your workflow grows past LLM-mediated research, the rest of the Form4API platform is right behind it:
- Webhooks — HMAC-signed, exponential backoff, dead-letter queue, auto-disable on persistent failure. For production pipelines, not just LLM chats.
- Python SDK —
pip install form4api(PyPI) - JS / TypeScript SDK —
npm install form4api(npm) - Dashboard — usage, billing self-serve, key rotation, webhook health, billing history.
The MCP wraps the same backend as all of the above — every fact your LLM cites can be re-fetched programmatically through any of these channels with the same key.
Plans
25 of the 39 tools work on the free plan, and every tool that is free today stays free. New premium capability gets tiered as it ships; nothing that already works on your key is taken away later.
| Tool | Free | Pro | Business |
|---|---|---|---|
search |
✓ | ✓ | ✓ |
get_transactions, get_recent_filings, get_filing |
✓ | ✓ | ✓ |
get_insider_profile, get_insider_transactions |
✓ | ✓ | ✓ |
get_company_overview, get_company_insiders, get_company_peers |
✓ | ✓ | ✓ |
get_insider_career_summary, get_insider_scorecard |
— | ✓ | ✓ |
get_insider_leaderboard, get_signals, get_sentiment |
— | — | ✓ |
get_form144, get_holdings, get_managers |
— | — | ✓ |
explain_signal, list_schedule13_dg |
— | — | ✓ |
list_congress_trades |
✓ (30-day disclosure window) | ✓ (unlimited history) | ✓ (unlimited history) |
list_congress_politicians, get_congress_politician, get_congress_ticker_rollup, get_convergence_signals |
— | ✓ | ✓ |
| Requests/day | 500 | 50,000 | 250,000 |
| Requests/minute | 60 | 200 | 500 |
get_transactions paging depth |
20 pages | unlimited | unlimited |
Starter sits between Free and Pro: 7,500 requests/day and 120/minute, 100 pages of get_transactions paging depth and a 366-day congressional history window. Tools marked Pro or Business in the table still need that plan. Enterprise is unlimited requests/day at 2,000/minute.
For a bulk historical pull, use the REST /v1/transactions/export endpoint (Business) rather
than paging — it streams the whole filtered set as CSV in one request.
What your agent sees at a paywall
A gated call never surfaces a bare HTTP error. The MCP returns a structured
upgrade_required payload so the model can explain the situation and the fix in one turn:
{
"error": "upgrade_required",
"required_plan": "business",
"current_plan": "Free",
"message": "This endpoint requires the Business plan or higher. Your current plan is Free.",
"unlocks": "The Business plan includes cluster-buy signals and sentiment scores, 13F institutional holdings and managers, Form 144 notices, bulk CSV export, and 250,000 requests/day."
}
message is the API's own explanation, kept verbatim — it names the specific limit or
parameter that stopped the call, which is usually what the model needs to suggest a working
alternative. The same shape is returned when a Pro-only parameter is used on an otherwise
free tool, so the model can simply retry without that filter.
upgrade_url is included only when the API supplies one; it is omitted otherwise.
Data coverage
- 1.38M+ Form 4 transactions from SEC EDGAR
- 620K+ filings across all reporting companies
- 125K+ Form 144 notice-of-proposed-sale filings (Business+)
- 45M+ Form 13F-HR holdings across 118K+ filings, $80T+ AUM in the latest complete quarter (Business+)
- 3.3+ years of history (since 2023-04)
- 10b5-1 plan flag on every transaction
- Amendment-aware — Form 4/A reconciled
- Congressional STOCK Act trades — U.S. House Clerk periodic transaction reports only (Senate eFD is not ingested).
list_congress_tradesis available on every plan with a history window by plan (Free 30 days, Starter 366 days, Pro+ unlimited); politician rollups and convergence need Pro+. Digital filings, amounts always shown as disclosed ranges (amountLow/amountHigh), never a fabricated midpoint, plusdisclosureLagDayson every trade (up to 45 days under the STOCK Act) - Real-time ingestion — new filings within minutes of SEC publication
Install as a Claude Desktop Extension (DXT)
A manifest.json is included at the repo root for one-click install via the Desktop Extensions (DXT) format. Once Claude Desktop supports .dxt files natively, pack and install with:
npx @anthropic-ai/dxt pack
# Produces form4api-mcp.dxt — open it in Claude Desktop to install
Until then, use the standard claude_desktop_config.json method described in Quick install above.
How tools stay in sync with the backend
This MCP is split between two layers:
- Hand-written tools in
src/tools/*.ts(transactions, signals, sentiment, form144, holdings, …) — these carry the LLM-discriminator descriptions (amendment-aware, 10b5-1 clean, etc.) that make this MCP pick correctly over alternatives. - Auto-generated tools in
src/tools/_generated.ts— produced fromhttps://api.form4api.com/openapi/v1.jsonbynpm run codegen. Every new backend endpoint that lands in the OpenAPI spec flows in here automatically. CI runsnpm run codegen:checkon every PR and fails the build if the committed file drifts from what the live spec would produce, so the MCP is never silently behind the backend.
To add a new generated tool: ship the endpoint on the backend, regenerate (npm run codegen), commit src/tools/_generated.ts, publish. No tool-wrapper code needed.
The 6 recipe prompts live in src/prompts/recipes.ts — also hand-written, not generated. They add no new backend surface area; each one is a client-side template that tells the LLM which existing tools to call and in what order.
Links
Form4API MCP questions
How do I add Form4API MCP to Claude Code?
Run claude mcp add --env 'FORM4API_KEY=<YOUR_FORM4API_KEY>' form4api-mcp -- npx -y form4api-mcp in your terminal. Add --scope user to make it available in every project.
How do I add Form4API MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What tools does Form4API MCP provide?
It exposes 40 tools, including get_transactions, get_recent_filings, get_filing, get_insider_profile, get_insider_transactions.
How can I verify if my MCP configuration is correct?
You can run the verify_setup tool, which confirms your API key is valid and the MCP server is reachable.
Can I try the server without an API key?
Yes, the get_public_stats tool is keyless and allows you to preview live data coverage before signing up.
What is the difference between this and other SEC APIs?
Unlike some competitors, this server provides both 13F institutional ownership and insider transaction data in a single response.