Since the STOCK Act (2012), members of the US Congress must disclose their personal stock trades within 45 days. Sites like Capitol Trades and QuiverQuant turned "Pelosi trades" into a retail phenomenon. But where does that data actually come from, and what does it take to use it programmatically?
The official sources
- House of Representatives: the House Clerk publishes Periodic Transaction Reports (PTRs) as PDFs on disclosures-clerk.house.gov. Newer reports are e-filed and machine-readable; older ones are scanned paper.
- Senate: the Senate eFD system - which sits behind an aggressive anti-bot wall and, importantly, prohibits commercial use of the data in its terms.
That asymmetry matters. Any commercial dataset that includes Senate trades is in murky territory. We made the deliberate choice to cover the House only, from the official Clerk feed, e-filed reports only (no OCR guessing on scans). Less coverage, clean provenance.
What parsing actually involves
A PTR PDF contains the member's transactions: the asset name and ticker, transaction type (purchase/sale), date, and an amount range (disclosures use brackets like $1,001-$15,000, not exact values). Turning that into data means extracting tables from PDFs, normalizing tickers, classifying buy vs sell, and estimating sizes from the brackets - then scoring recency and size so a $500K purchase last week ranks above a $2K sale from a month ago.
What the feed returns
The US Congress Trading Monitor takes House member names and returns their recent trades parsed from the official feed: ticker, side, estimated size, date, and a recency/size score, with each record traceable to the underlying disclosure. It costs $0.20 per member analyzed - against QuiverQuant's $25/month subscription, that is the difference between paying for browsing and paying for data you actually pull. (Their dashboard is genuinely good for daily browsing; see our comparison.)
For pipelines: JSON via API, webhooks on every run, or let an AI agent query it through the free MCP server alongside insider clusters and 8-K events - "did any House member trade a stock that also had an insider buying cluster this month?" is a three-tool agent query.
The honest caveats
Disclosures arrive up to 45 days after the trade, amounts are ranges rather than exact figures, and academic evidence on copying congressional trades is mixed at best. This is transparency data - useful for research, journalism and screening, not a trading system.
Data for research, screening and monitoring - not investment advice.
Prefer the tool to the write-up? Browse all data feeds or connect the free MCP server.