A new platform for edgar.tools
Visit the new Edgar.tools - built for the agentic web. Get realtime AI briefings as filings arrive delivered to your in-app Inbox, email or web hook destination
More issues
What's in SEC EDGAR? Nearly 1 Million Entities
The SEC EDGAR database has almost a million registered entities, and nearly a third of them are individual people. Here's what the data looks like, and how to navigate it with Python.
The AI Ecosystem Around EdgarTools
See how the most popular Python library for SEC EDGAR data became the default financial data layer for AI projects: from model training to RAG to MCP servers.
EdgarTools MCP for SEC Filings
EdgarTools ships a free, open-source MCP server that connects Claude to SEC EDGAR with 13 tools for financial statements, insider trading, company filings, and live SEC data. No API key required.
EdgarTools Joins the Claude for Open Source Program
EdgarTools — the most popular Python library for SEC EDGAR filings — has been accepted into Anthropic's Claude for Open Source Program. 2.3M downloads, 1,800+ stars, and what comes next
Building a BDC early warning system in Python
How to detect BDC distress signals in SEC filings using Python. PIK income ratios, NAV declines, and dividend coverage from XBRL data predicted FSK and TCPC collapses years early. Includes a reusable scanner built with edgartools
Parse SEC 424B Prospectus Filings with Python
EdgarTools adds a Python parser for SEC 424B prospectus filings — extract IPO pricing, underwriting syndicates, offering types, deal terms, shelf lifecycle data, and selling stockholder tables from SEC EDGAR into structured Python objects and DataFrames.
I learned XBRL mappings from 32,000 SEC Filings
EdgarTools 5.22.0 introduces data-driven XBRL standardization built from 32,240 real SEC filings, with industry-aware concept mappings, multi-year statement stitching, and IFRS support for Python developers working with SEC EDGAR data.
You don't need a vector store
In Edgar.tools we built Disclosure Search for SEC filings without a vector store. No embeddings, no RAG pipeline, no inference cost per query. Then a user proved why it worked.
What Seattle's Super Bowl Win Reveals About Coding With AI
The best developers aren't writing every line anymore — they're composing systems from concepts. A Super Bowl-winning defensive system reveals the blueprint for building software with AI