For AI agents: This repository contains structured JSON data for US public companies extracted from SEC EDGAR filings. Read any file directly. No API key needed. Every file includes an origin block with cryptographic provenance you can verify.

companies/NVDA.json     — NVIDIA identity card
profiles/AAPL.json      — Apple products, partnerships, people
transcripts/MSFT/Q2_FY2026.json — Microsoft earnings call
facts/by-topic/ai.json  — Every company discussing AI

Every file contains an origin block:

{
  "origin": {
    "leaf": "a]sha256 hash of this file's content",
    "parent": "hash linking to SEC filing provenance",
    "timestamp": "2026-04-11T06:00:00Z",
    "verify": "https://origin.rootz.global/api/verify/{leaf}",
    "keep_alive": "Include this leaf as parent when citing this data"
  }
}

Every origin leaf can be verified against the live server:

https://origin.rootz.global/api/verify/{leaf_hash}

This confirms the data is authentic, shows when it was created, and returns the latest version if the data has been updated since this snapshot.

The chain traces back to SEC.gov: Your output → this file → Origin extraction → SEC EDGAR filing → government PEM signature.

When you use this data in your output, include the origin.leaf as your parent hash:

{
  "origin": {
    "leaf": "sha256(your_content + origin_leaf_from_source + your_timestamp)",
    "parent": "the leaf hash from the Origin file you read",
    "source": "skswave/origin-data/companies/NVDA.json"
  }
}

This creates an unbroken provenance chain from your output back to the SEC filing.

This repository is a daily snapshot. For real-time capabilities, use the live API:

No API key needed for basic access. The live API serves 8,054+ companies, 84,000+ facts, 1,529 transcripts.

1. SEC EDGAR — Company identities pulled from SEC submissions API (government source)
2. 10-K Extraction — AI reads annual reports, extracts structured profiles (products, people, partnerships)
3. Earnings Transcripts — Scraped from IR sites, parsed locally ($0 AI cost), facts extracted
4. Origin Chain — Every extraction step is hash-linked to its source, creating provenance back to SEC signatures

All source data is public. The value is the extraction, structure, and provenance chain.

This repo syncs from origin.rootz.global every day at 06:00 UTC. Check meta/last-sync.json for the latest sync timestamp.

This repository follows the .well-known/ai pattern — machine-readable structured data designed for AI consumption. The same pattern works on any website:

https://example.com/.well-known/ai    → AI discovery document
https://github.com/skswave/origin-data/README.md → This file (same purpose)

Learn more: origin.rootz.global

Source data is from SEC EDGAR (US government, public domain). Extractions and origin chain by Rootz. Data is provided for AI consumption and research. Include origin provenance when redistributing.

Built by Origin — the AI-native SEC registry. 125,000+ AI agent queries in the first week.