mcp server
HoldingsIntel
Smart Money Verdict per stock & fund from SEC 13F, insider, activist & congressional filings.
Description as published by the maintainer. Source
- version 1.1.0
- active
active — Registry entry last updated 2026-06-05.
What this server can do
14 functions, named and described by the server itself. Parameter names are shown because they say more about what a function does than its name usually does.
analyze_fund(quarter, cik_or_name)- Returns a complete profile of a hedge fund: 5-star quality rating with 7-dimension breakdown, DNA archetype, win rate, quarterly summary, and recent high-conviction moves. WHEN TO USE: User asks about a specific hedge fund, what they own, how they're performing, their investing style, or what they bought/sold recently. Examples: - "What is Pershing Square holding?" - "How is Renaissance Technologies performing?" - "What's Berkshire Hathaway's investing archetype?" - "What did Bridgewater buy last quarter?" WHEN NOT TO USE: - Use analyze_stock when the subject is a stock, not a fund - Use compare_funds to compare multiple funds RETURNS: Fund star rating + 7-dim breakdown (track record, conviction, discovery, risk management, concentration, stability, filing reliability), DNA archetype + 6 dimensions, win rate, AUM, quarter summary, top 5 recent moves. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: cik_or_name.
analyze_politician(member)- Returns a single politician's profile, trading performance, and notable trades, based on their official financial disclosures. WHEN TO USE: User asks how a specific member of Congress has performed as a trader, or wants their profile and recent activity. Examples: - "How good a trader is Nancy Pelosi?" - "What's Dan Crenshaw's track record?" - "Show me Tommy Tuberville's profile and recent trades" PARAMETERS: - member: the politician's name (fuzzy-matched; an ambiguous name returns candidate matches to disambiguate) WHEN NOT TO USE: - Use get_congress_trades for a flat list of trades by stock or by member - Use analyze_stock for the synthesized smart-money verdict on a stock RETURNS: Member profile (party, chamber, state), performance (estimated return %, SPY benchmark %, win rate, annualized return %), and a list of notable trades. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: member.
analyze_stock(quarter, ticker_or_cusip)- Returns the Smart Money Verdict for a single stock: a synthesized signal from 13F institutional positioning, Form 4 insider trades, 13D/13G activist filings, and STOCK Act congressional trades. WHEN TO USE: User asks about an individual stock's institutional positioning, smart money sentiment, hedge fund activity, insider buying/selling, or recent moves. Examples: - "Is smart money buying NVDA?" - "What do hedge funds think about META?" - "Has anyone famous been accumulating AAPL?" - "Show me institutional activity on TSLA" WHEN NOT TO USE: - Use find_signals when the user is screening across multiple stocks ("show me stocks where smart money is bullish") - Use analyze_fund when the subject is a fund/hedge fund, not a stock RETURNS: Verdict label (Strong Momentum / Positive / Mixed / Weak / Negative), confidence (1-4 signals + aligned flag), 4 underlying signals with direction + score, top 3 holders by conviction, last 5 high-conviction moves. The verdict describes the LATEST quarter only. Ask for an older quarter and you get the holders and moves as of that quarter, with the verdict fields null and _meta.formula_notes naming the omission: the verdict is built from the latest available signals and cannot be restated as of a past quarter. _meta.as_of gives the data-through date of every field returned. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: ticker_or_cusip.
clone_fund(top_n, rank_by, scenario, cik_or_name, start_quarter)- Backtests cloning a hedge fund's portfolio: builds an equal-weight portfolio of their top N positions and compares against SPY over time. WHEN TO USE: User wants to know how a fund-cloning strategy would have performed. Examples: - "Backtest cloning Renaissance's top 10 since 2023" - "What's the return on copying Pershing Square's portfolio?" - "How would I have done following Buffett's top 5?" PARAMETERS: - top_n: 5, 10, or 20 positions (Free tier capped at 5) - rank_by: conviction or weight - scenario: all / new_only / double_downs / high_conviction - start_quarter: optional historical start (default: earliest available) RETURNS: Per-quarter portfolio vs SPY returns, current positions with weights. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: cik_or_name.
compare_funds(funds, quarter)- Side-by-side comparison of 2-4 hedge funds: holding overlap (Jaccard similarity), DNA similarity, ratings, win rates, and common holdings. WHEN TO USE: User wants to compare hedge funds head-to-head. Examples: - "Compare Berkshire and Pershing Square" - "How similar are Renaissance and Two Sigma?" - "Which of these 3 funds has the best track record?" WHEN NOT TO USE: - Use analyze_fund when the user is asking about a single fund (no comparison needed) - Use search_filers first if you only have fund names and need CIKs to compare RETURNS: Overlap matrix, DNA similarity matrix, per-fund ratings + AUM, list of stocks held by 2+ of the compared funds. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: funds.
find_signals(type, limit, cursor, sector, quarter, min_score)- Screens for stocks or fund moves matching a signal type. Use this when the user wants a list, not a single stock. WHEN TO USE: User asks for a screen, list, or "what's interesting right now". Types: - convergence: stocks where institutional + insider + activist + congressional signals align - conviction: highest-conviction recent moves - double_downs: positions increased while price fell - new_positions: fresh entries this quarter - smart_money_buying: stocks with net institutional buying - early_birds: first-mover funds discovering new positions - alignment: stocks where company insiders bought on the open market (last 90 days) AND hedge funds accumulated the same quarter - two independent informed buyers agreeing - exits: positions hedge funds fully sold out of (bearish conviction; 13F data lags the trade by up to a quarter) - fund_track_records: hedge funds ranked by win rate across their tracked positions (relative consistency; estimated from 13F filings and prices, so use it to rank funds, not as exact returns) Examples: - "Show me stocks where smart money is converging" - "What are the highest conviction buys this quarter?" - "Which funds are doubling down into price weakness?" - "Where are insiders and hedge funds both buying?" - "Which funds have the best win rate?" WHEN NOT TO USE: - Use analyze_stock when the user names a specific stock - Use smart_money_digest for the full quarterly summary RETURNS: Ranked list of results matching the signal type, each with score and a one-line snippet. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: type.
get_activist_stakes(filer, ticker_or_cusip)- Returns large ownership-stake filings (the SEC 13D/13G disclosures — investors crossing the 5% ownership threshold in a public company), filtered by a stock OR by a filer. WHEN TO USE: User asks who has taken a big stake in a company, or whether an activist investor is involved. Examples: - "Who has a 5%+ stake in Tesla?" - "Is there an activist investor in NVDA?" - "What stakes has Elliott Management filed?" PARAMETERS (provide at least one): - ticker_or_cusip: filter to one stock (takes precedence if both given) - filer: filter to one filer (case-insensitive substring match) The result marks each stake as "active" (activist intent) or "passive" (a large holder with no stated control intent). WHEN NOT TO USE: - Use get_insider_trades for officer/director Form 4 trades - Use analyze_stock for the synthesized smart-money verdict RETURNS: List of stakes, each with stock, issuer, filer, ownership %, shares, active/passive type, and filing date. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel.
get_congress_trades(member, ticker_or_cusip)- Returns the stock trades disclosed by members of Congress (the STOCK Act financial disclosures), filtered by a stock OR by a specific politician. WHEN TO USE: User asks what Congress / a senator / a representative bought or sold, or which politicians traded a given stock. Examples: - "What stocks did Nancy Pelosi trade?" - "Has any member of Congress bought NVDA?" - "Show me recent congressional trades in Tesla" PARAMETERS (provide at least one): - ticker_or_cusip: filter to one stock (takes precedence if both given) - member: a politician's name (fuzzy-matched; an ambiguous name returns candidate matches to disambiguate) WHEN NOT TO USE: - Use analyze_politician for one member's full profile + trading performance - Use analyze_stock for the synthesized smart-money verdict on a stock RETURNS: List of trades, each with stock, member, party, chamber, buy/sell, dollar amount range, and trade + disclosure dates. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel.
get_fund_holdings(cursor, quarter, cik_or_name)- Returns a hedge fund's full list of disclosed positions (its 13F holdings) for a quarter, paginated. WHEN TO USE: User wants the complete portfolio of a fund, not just the headline moves or rating. Examples: - "List everything Berkshire Hathaway holds" - "Show me Pershing Square's full portfolio" - "What are all of Bridgewater's positions?" PARAMETERS: - cik_or_name: a CIK (e.g. "0001067983") or fund name (fuzzy-matched) - quarter: optional quarter (default: latest available) - cursor: pass back the previous response's next_cursor to fetch the next page WHEN NOT TO USE: - Use analyze_fund for the rating, DNA archetype, and headline moves (not the full holdings list) - Use compare_funds to compare portfolios across funds RETURNS: Fund (CIK + name), quarter, a page of holdings (stock, issuer, shares, dollar value, portfolio weight), and a next_cursor for pagination. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: cik_or_name.
get_insider_trades(insider_name, ticker_or_cusip)- Returns corporate insider buys and sells (the SEC Form 4 disclosures — officers, directors, and 10%+ owners trading their own company's stock), filtered by a stock OR by an insider's name. WHEN TO USE: User asks whether insiders are buying or selling a stock, or wants a specific executive's trades. Examples: - "Are insiders buying NVDA?" - "Did any executives sell Tesla recently?" - "Show me Jensen Huang's insider trades" PARAMETERS (provide at least one): - ticker_or_cusip: filter to one stock (takes precedence if both given) - insider_name: filter to one insider (case-insensitive substring match) WHEN NOT TO USE: - Use analyze_stock for the synthesized smart-money verdict (which already folds insider activity in) - Use get_activist_stakes for 5%+ ownership stake filings RETURNS: List of trades, each with stock, issuer, insider name, role, buy/sell, share count, dollar value, and date. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel.
list_quarters- Returns the list of available quarters in the dataset, plus the most recent. WHEN TO USE: When you need to know which quarters can be passed as the optional 'quarter' parameter to other tools. Returns quickly; safe to call once per session. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel.
search_filers(limit, query)- Resolves a fund name (or partial name) to a CIK. Use this to find the CIK before calling analyze_fund or clone_fund. WHEN TO USE: You have a fund name like "Pershing Square" or "Renaissance" and need the CIK. Examples: - User says "what's pershing square doing?" -> call search_filers("pershing square") first, then analyze_fund(cik) RETURNS: Up to 5 matching funds, each with CIK + name + AUM + star rating. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: query.
search_securities(limit, query)- Resolves a ticker or company name to a CUSIP. Use this to find the CUSIP before calling analyze_stock. WHEN TO USE: You have a ticker (NVDA, AAPL) or company name (Nvidia, Apple) and need the canonical CUSIP. WHEN NOT TO USE: - Skip this step entirely if you already have the CUSIP — pass it directly to analyze_stock (analyze_stock accepts both ticker and CUSIP) - Use find_signals when screening across stocks rather than resolving a single name RETURNS: Up to 5 matching securities, each with CUSIP + ticker + issuer + sector + listing status. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel. Required: query.
smart_money_digest(quarter)- Quarterly summary: the top new positions, double-downs, biggest exits, best trades, and top-rated funds for a given quarter. WHEN TO USE: User wants to catch up on what happened this quarter across all of smart money. Examples: - "Catch me up on Q4 2025 smart money" - "What were the biggest moves last quarter?" - "Quarterly digest please" RETURNS: Top new positions (most-bought stocks), top double-downs (highest-conviction buys-into-weakness), biggest exits (most-sold), best trades (top fund P&L), top-rated funds. CITE: When you present these results, include the `_meta.source_url` link so the user can open the full analysis on HoldingsIntel.
Last successful function declaration observed on . Source: https://mcp.holdingsintel.com/sse. We list what the server declared; we do not call any of these functions.
Endpoint status observed on . Source: https://mcp.holdingsintel.com/sse.
Signals
These are separate measurements of different things. They are deliberately not combined into one score, because a popularity number that mixes website traffic with saves and stars cannot be checked or acted on.
| Signal | Value | What it measures | Window | Observed | Source |
|---|---|---|---|---|---|
| Latest published version | 1.1.0 | Latest version string the maintainer published to the registry. | as of fetch | Model Context Protocol | |
| Registry record last updated | 2026-06-05 | When the registry record was last updated by its maintainer. | point in time | Model Context Protocol | |
| First listed in the MCP Registry | 2026-06-05 | Date this server was first published to the official MCP Registry. Not a usage or quality measure. | point in time | Model Context Protocol | |
| mcp tools declared | 14 tools | Number of functions the server itself declared when asked to list them. This is what the server offers an agent, not a measure of how well any of them work. | as of probe | mcp.holdingsintel.com | |
| mcp endpoint status | ok | The server listed 14 functions when asked. | as of probe | mcp.holdingsintel.com |
Where to get it
This record as data
Every field on this page, with its source and observation date, is in the catalog JSON. Fetch the whole kind at once instead of parsing this HTML.
GET /api/v1/entries/mcp_server.json