edgar-mcp

SEC filings, inside your agent.

An MCP server that lets an LLM resolve companies, browse new securities offerings, screen the actual deal economics, and read insider trades — straight from the SEC. All read-only, all public data, no API key.

See it in action

One prompt in Claude Code, start to finish:

Using edgar, what new Reg D raises were filed in California recently? Pick the biggest and tell me how much they're raising and who's behind it.

Claude Code using edgar-mcp to find recent California Reg D raises and screen the biggest one

What happened

Claude called get_recent_offerings(form="D", state="CA") — a fresh list of California Reg D filings, newest first.
It screened the candidates with get_form_d_details, pulling each offering's size, amount sold, exemptions, and the people behind it.
It surfaced the standout: DCVC Energy & Climate II, L.P. — a $500M deep-tech climate fund, brand-new (first sale "yet to occur"), exempt under Rule 506(c) + 3(c)(7), run by DCVC co-founders Zachary Bogue and Matthew Ocko.
…then ranked the rest of the batch — Restora Medical ($15M), U&I Ventures ($25M), 8ALPHA AI, Planar — for context.

Every figure is pulled live from SEC filings through the tools — not the model's memory.

What you can ask

Real exchanges, driven entirely by the tools below.

Who runs Reddit, and what have its insiders been trading?

Resolves Reddit, Inc. (RDDT), returns the leadership roster — co-founder Steve Huffman plus the COO, CFO, and Chief Legal Officer — and recent Form 4 insider transactions with the shares, price, and resulting stake (e.g. the COO's 38,112-share disposition at $154.88).

What new Reg D raises were filed in California this week?

Browses recent Form D filings filtered to CA, newest first — issuer, form, date, and a link — ready to screen.

For that raise, how much are they raising and who's behind it?

Parses the Form D: total offering, amount sold and remaining, minimum investment, number of investors, industry, revenue range, security type, exemptions, and the officers / directors / promoters — structured, not guessed.

Summarize Apple's latest 10-K risk factors, and show its headline financials.

Opens the filing, returns the document text (HTML stripped, paginated) to summarize, and pulls XBRL financials — FY revenue, net income, assets, equity, cash.

Tools

Tool	What it does
`lookup_issuer`	Name/ticker → CIK. Public and private / non-exchange filers.
`list_filings`	An issuer's recent filings, newest first; optional form filter.
`search_filings`	Full-text search across filing documents.
`get_recent_offerings`	Recent Reg CF (C) / Reg D (D) / Reg A (1-A) raises; optional state filter.
`get_filing`	Open a filing: metadata + every document with links.
`get_filing_text`	A document's text (HTML stripped), paginated — for reading/summarizing.
`get_form_d_details`	Parse a Reg D raise: amount, sold, investors, industry, people.
`get_form_c_details`	Parse a Reg CF raise: target/max, price, terms, two-year financials.
`get_company_facts`	XBRL headline financials for public companies.
`get_insider_trades`	Recent Form 4 transactions: owner, action, shares, price.
`get_insiders`	A company's insiders (officers, directors, >10% owners) and roles.

Install

Requires Python 3.12+ and uv.

Claude Code

claude mcp add edgar -- uvx mcpwright-edgar

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "edgar": { "command": "uvx", "args": ["mcpwright-edgar"] }
  }
}

OpenAI & other MCP clients

It's a standard stdio MCP server — so it works beyond Claude: the OpenAI Agents SDK, Cursor, VS Code, Cline, Goose, Zed, and more all launch it the same way.

{
  "mcpServers": {
    "edgar": { "command": "uvx", "args": ["mcpwright-edgar"] }
  }
}

An OpenAI Agents SDK (Python) example is in the README. (Hosted chat connectors like ChatGPT want a remote HTTP server; this runs locally over stdio.)

The SEC asks for a descriptive contact in the User-Agent — set EDGAR_MCP_USER_AGENT="your-app you@example.com". The server throttles, retries, and caches for you.

Source, full tool reference, and roadmap → github.com/mcpwright/edgar-mcp

Questions or ideas? → Discussions · found a bug? → Issues