Agent-ready alternative data, for indie traders.
Plug HedgeFriend into Claude, Cursor, or any MCP-compatible client with one API key. Your agent can reach every HedgeFriend endpoint — SEC filings, insider trades, congressional disclosures, lobbying, layoffs, hiring, macro, and news sentiment — and finds the right one by asking. No glue code, no schemas to memorize, no scraping.
> Why did $NVDA sell off yesterday? Check insiders, congress, and news sentiment. ⏺ tool find_data (3 endpoints matched, free) ⏺ tool batch_requests (3 calls, 1 round-trip) Three signals stand out: 1. Insiders were net sellers of $12.4M over the past 90 days across 8 distinct officers. 2. Two Senate members disclosed sales in the $15K–$50K range on 08/22 and 08/24. 3. Sentiment turned bearish (-0.31 avg, article count 3.2× baseline) after a Reuters piece on hyperscaler capex guidance.
Skip the glue code. Your model calls the tools directly.
Pasting raw JSON into a prompt is not a research workflow. MCP lets your agent discover endpoints, batch related calls, and cite the exact tool and parameters it used — the way a human analyst would.
Same HedgeFriend API key you'd use with curl. Works with Claude Desktop, Claude Code, Cursor, and anything else that speaks MCP.
Your agent describes what it wants and find_data returns the endpoints that answer it — searching the catalog costs nothing. batch_requests then fans 20 calls out concurrently in one round-trip, so multi-signal questions stay fast.
No parallel auth story to configure. The MCP forwards your bearer token straight to the API; your daily quota is the same one you see in the dashboard.
Three steps. Under a minute.
- 1Grab your API key
Free tier is live at /signup — 500 requests/day, all endpoints. No card required.
- 2Paste one config block
Add the JSON below to your client's MCP config file. Same shape for Claude Desktop and Cursor; one shell command for Claude Code.
- 3Ask your agent anything
Restart the client. The HedgeFriend tools show up automatically. Try one of the example questions below.
{
"mcpServers": {
"hedgefriend": {
"url": "https://mcp.hedgefriend.dev/mcp/",
"headers": { "Authorization": "Bearer hf_..." }
}
}
}{
"mcpServers": {
"hedgefriend": {
"url": "https://mcp.hedgefriend.dev/mcp/",
"headers": { "Authorization": "Bearer hf_..." }
}
}
}claude mcp add hedgefriend \ --transport http \ --header "Authorization: Bearer hf_..." \ https://mcp.hedgefriend.dev/mcp/
- Is NVDA insider activity net-buying or net-selling in the last 90 days?
- Which S&P 500 tickers had congressional purchases before an FDA action?
- How has SNAP's engineering headcount changed since the last 10-Q?
- Show me any WARN filings from Tesla suppliers in the last month.
- Correlate PLTR Wikipedia pageviews with news sentiment over 6 months.
- What did the most recent DEF 14A say about ORCL CEO comp?
Every endpoint, one bearer token.
Rather than one tool per endpoint, the server gives your agent a way to search the catalog and a way to call anything in it — so every route in the docs is reachable, including ones added after you set the client up. Shortcuts cover the handful of calls almost every session starts with.
- find_dataAsk a question in plain English, get the endpoints that answer it
- describe_dataParameters, trading signal, and a worked example for one endpoint
- get_catalogThe whole endpoint directory in one call
- call_apiCall any endpoint in the catalog — the general-purpose tool
- batch_requestsFan out up to 20 calls in one round-trip
- search_symbolsResolve a company name to a ticker
- get_sec_filings10-K / 10-Q / 8-K / Form 4 by ticker
- get_insider_tradesIndividual Form 4 transactions
- get_financialsXBRL fundamentals (Revenue, NetIncome, EPS, …)
- get_congressional_tradesSTOCK Act House/Senate disclosures
- get_news_sentimentDaily sentiment time series
- get_usageCheck your daily quota and remaining requests
- submit_feedbackFlag stale, missing, or wrong data for triage
Give your model a research desk.
Free tier is live. One key, one config block, and Claude can pull insider trades, congressional disclosures, and news sentiment for any ticker.