I've been building StockFit API for about a year — a REST API that turns SEC
EDGAR filings into clean JSON you can actually use.
The backstory: I kept starting side projects that needed financial data and
kept burning 80% of my time just wrestling with EDGAR instead of building.
XBRL is technically public and free, but "free" is doing a lot of work there
— tagging is inconsistent across filers, fiscal years drift, Q4 isn't even a
standalone period (you derive it from the 10-K minus the 9-month Q3 figure or the next best available data point),
and the same financial concept maps to half a dozen different tags. Every
fintech dev I know has written and thrown away their own EDGAR parser at
least once.
So I built the thing I wanted. Under the hood:
- Arelle for XBRL processing
- DuckDB for the fact store (~250M facts, ~5M filings)
- Daily ingestion from EDGAR, with Q4 reconstruction at read time
- Curated fact mapping on top of raw XBRL concepts so you don't have to know
that us-gaap:Revenues, us-gaap:SalesRevenueNet, and
us-gaap:RevenueFromContractWithCustomerExcludingAssessedTax all mean
roughly the same thing depending on the filer
- REST API with OpenAPI docs and an MCP server if you want to use it from
Claude/Cursor/etc
Coverage right now: financials, ownership, ETF/mutual fund exposure, insider
transactions (Forms 3/4/5), institutional 13F holdings, executive comp from
proxy filings, the raw filings themselves, and screeners for Stocks/ETFs.
The free tier is 100 requests/day, no credit card. Paid tiers start around
$19/month if you need more.
Things I'd genuinely value feedback on:
1. The API shape — I made a bunch of opinionated decision. Happy to
change things if they could be better.
2. The Q4 reconstruction logic — it works, but I've seen edge cases with
non-December fiscal year ends (AAPL=Sep, COST=Aug) - I'm reasonably sure I've
covered all of them.
3. The docs - are they good the way they are
Not things I'd love feedback on (I know, they're on the roadmap):
- Historical point-in-time data earlier than ~2010 is spotty
- Real-time quotes — this is fundamentals data, not a market-data feed
- International filings — US-only for now
I've been building StockFit API for about a year — a REST API that turns SEC EDGAR filings into clean JSON you can actually use.
The backstory: I kept starting side projects that needed financial data and kept burning 80% of my time just wrestling with EDGAR instead of building. XBRL is technically public and free, but "free" is doing a lot of work there — tagging is inconsistent across filers, fiscal years drift, Q4 isn't even a standalone period (you derive it from the 10-K minus the 9-month Q3 figure or the next best available data point), and the same financial concept maps to half a dozen different tags. Every fintech dev I know has written and thrown away their own EDGAR parser at least once.
So I built the thing I wanted. Under the hood:
- Arelle for XBRL processing - DuckDB for the fact store (~250M facts, ~5M filings) - Daily ingestion from EDGAR, with Q4 reconstruction at read time - Curated fact mapping on top of raw XBRL concepts so you don't have to know that us-gaap:Revenues, us-gaap:SalesRevenueNet, and us-gaap:RevenueFromContractWithCustomerExcludingAssessedTax all mean roughly the same thing depending on the filer - REST API with OpenAPI docs and an MCP server if you want to use it from Claude/Cursor/etc
Coverage right now: financials, ownership, ETF/mutual fund exposure, insider transactions (Forms 3/4/5), institutional 13F holdings, executive comp from proxy filings, the raw filings themselves, and screeners for Stocks/ETFs.
The free tier is 100 requests/day, no credit card. Paid tiers start around $19/month if you need more.
Things I'd genuinely value feedback on:
1. The API shape — I made a bunch of opinionated decision. Happy to change things if they could be better. 2. The Q4 reconstruction logic — it works, but I've seen edge cases with non-December fiscal year ends (AAPL=Sep, COST=Aug) - I'm reasonably sure I've covered all of them. 3. The docs - are they good the way they are
Not things I'd love feedback on (I know, they're on the roadmap):
- Historical point-in-time data earlier than ~2010 is spotty - Real-time quotes — this is fundamentals data, not a market-data feed - International filings — US-only for now
Link: https://developer.stockfit.io Docs: https://api.stockfit.io/docs
Happy to answer questions or chat about the project if anyone's interested in building on top of it. Thanks for reading!