EdgarTools Joins the Claude for Open Source Program

Last week I received an email from Anthropic:

EdgarTools has been accepted into Anthropic's Claude for Open Source Program, which gives qualifying open-source maintainers six months of free access to Claude Max 20x — Anthropic's highest-tier plan at $200/month.

This felt like a good moment to step back and share where the project stands.

The Numbers

EdgarTools started in December 2022 as a weekend project to make SEC Edgar data less painful to work with in Python. Three years later:

Who's Using It

The intelligence reports I run on the project tell an interesting story about who is adopting. In the last six weeks alone:

Finance professionals now make up 31% of notable new stargazers. Hedge fund managers, fintech infrastructure engineers, and quantitative analysts are showing up in the star log. The founder of Hedge Fund Alpha starred the repo. An engineer from moov-io (fintech payments infrastructure, 316 GitHub followers) starred it. These aren't casual drive-bys.

AI engineers are the fastest-growing cohort. A YC W25 company building AI infrastructure. Data engineers building LLM pipelines. The MCP Server integration — which lets Claude and other AI agents query SEC filings directly — is pulling in a new class of user that didn't exist a year ago.

Academic researchers from Stanford and Columbia have started appearing. When PhD students in CS departments start using your financial data library, something is working.

And then there's the number that surprised me most: 803 organizational forks. That's 2.5x the public fork count. It means teams inside companies are cloning the repo into private infrastructure. EdgarTools is running inside organizations I've never heard from.

What Got Us Here

The program requires maintainers of projects with 5,000+ GitHub stars or significant ecosystem impact. EdgarTools qualifies under the ecosystem impact path — a specialized library that 218 downstream projects depend on, with nearly half a million monthly downloads and active institutional deployment.

But the real answer to "what got us here" is stubbornness about a single idea: SEC data should be as easy to work with as any other Python data source. Five lines of code to get Apple's income statement. One function call to parse XBRL. A single Company("AAPL") entry point that just works.

from edgar import Company

company = Company("AAPL")
financials = company.get_financials()
print(financials.income_statement)

That simplicity hides real complexity. XBRL taxonomies are a maze. Financial statement "stitching" — reconstructing multi-period statements from tagged data — is genuinely hard. The 65% of open issues that are data quality bugs aren't a sign of poor engineering; they're a sign that users are pushing the library into production with real money on the line.

What This Means for the Project

Claude Max 20x means I can use Claude Code — which I already use heavily for development — without worrying about usage limits. For a solo maintainer shipping 3,000+ commits a year across XBRL parsing, MCP integrations, and financial statement rendering, that matters.

More concretely, here's what I'm working on:

• Financial statement accuracy — the stitching engine that reconstructs balance sheets, income statements, and cash flow statements from XBRL data is the core value proposition. Making it work correctly across the full diversity of SEC filers is the hardest problem in the library.
• MCP Server — AI agents that can query SEC filings, extract financial data, and analyze company disclosures. This is where the most interesting new adoption is coming from.
• International filings — 20-F and 6-K support for foreign private issuers. About 15-20% of SEC submissions come from non-US companies.

A Thank You

To the 1,861 people who've starred the repo. To miruddfan, who files detailed stitching bugs that make the library better. To the 218 projects that depend on edgarTools. To the hedge fund analysts and Stanford PhD students and YC founders who are building on top of it.

And to Anthropic, for recognizing that open-source infrastructure — even in a niche like SEC filing analysis — matters.

If you're working with SEC data in Python, give it a try: