EdgarTools

What EdgarTools does

EdgarTools MCP server gives AI agents access to SEC EDGAR filings through the EdgarTools library. The underlying library converts filings into typed Python objects and DataFrames covering financial statements, insider transactions, institutional holdings, fund portfolios, proxy statements, 8-K events, and other SEC filing types. It also supports filing text, section extraction, markdown output, full-text search, filing discovery, and the live filing feed.

The server is suited to research and data workflows that need more than raw SEC JSON. Agents can retrieve standardized statement rows across companies, inspect discrete quarters or trailing-twelve-month figures, find filings by ticker, CIK, or accession number, and work with structured objects for forms such as 13F, Form 4, N-PORT, and 8-K.

How it works

EdgarTools runs against SEC EDGAR data and uses the library's parsing and normalization features before returning results to an MCP client. Filing results can be accessed as structured objects, DataFrames, text sections, or complete filings rendered as markdown for retrieval-augmented generation.

Search can span filing content dating back to 2001, with filters such as form type, filing date, and 8-K item. Filing retrieval also supports the broader EDGAR archive, including filings published since 1994. Financial statement helpers provide standardized labels, while quarterly views derive Q4 figures from annual filings where applicable.

Setup and configuration

The underlying EdgarTools package requires Python 3.11 or newer. Install it with uv add edgartools or pip install edgartools; the package is named edgartools, while its Python module is edgar.

Every SEC request must identify the caller with a contact email. In Python, configure this with set_identity("your.name@example.com"). EDGAR access does not require an API key, account, or paid rate-limit tier.

Tools and capabilities

• Search the full text of SEC filings with form, date, and item filters.
• Find filings by ticker, CIK, accession number, or other identifiers.
• Retrieve current filings from the SEC feed.
• Read complete filings as markdown or extract numbered sections such as Item 1A and Item 7.
• Query standardized income statements, balance sheets, and cash-flow statements across multiple periods.
• Inspect 13F institutional holdings, Form 4 insider transactions, N-PORT fund portfolios, and 8-K events or earnings releases.
• Calculate trailing-twelve-month revenue and retrieve discrete quarterly financial data.

Limitations and notes

The service depends on SEC EDGAR availability and access rules. SEC fair-access limits allow up to 10 requests per second, and EdgarTools is designed to stay within that limit. The library is MIT-licensed and runs in the user's own Python process. A hosted product, edgar.tools, is separate from the local library and should not be confused with this server.

For a single known filing or JSON document fetched once from a script, direct EDGAR access may be more appropriate than using EdgarTools. The local library is intended for repeatable workflows such as pipelines, backfills, backtests, model corpora, notebooks, and custom joins.