Free Humanoid Corpus

A public-domain prior art commons covering robots, humanoids, and humanoid-adjacent entities from private companies, open source repositories, science fiction, and academia.

Browse the corpus Cross-cuts GitHub

DOI License: CC0-1.0


The corpus IS the prior art commons. Every entry, at the moment of timestamped commit and quarterly release, is a defensive publication. There is no separate commons layer above this catalog — the structured data here, combined with the cryptographic timestamping in releases/, is itself the invalidity-contention material.


What it covers

Corpus Count Role
Private 57 Patent-thicket holders. Reference points for what’s been claimed.
Academic 27 Peer-reviewed disclosures. Already prior art automatically.
Fictional 11 Science fiction. Cited successfully as 102 prior art when specific.
Open 11 Existing open hardware. Already public-domain or permissive.
Total 106 (77 commons-grade, 29 draft)

Entries span 1920 (Karel Čapek’s R.U.R.) through 2025.


The 48 cross-cuts are the working tool

Each subsystem cross-cut is a chronologically-ordered list of every disclosing entry, with prior_art_notes and disclosure_citation ready to drop into an invalidity contention.

All 48 cross-cuts


Provenance

Quarterly releases carry three independent cryptographic timestamps attesting pre-existence:

  • FreeTSA (RFC 3161)
  • DigiCert (RFC 3161)
  • OpenTimestamps (Bitcoin-anchored)

The current release tarball, corpus-2026.Q2.tar.gz (SHA-256 aa9430c6e785a409e3dbb10042b16e0e5677752c85eeffcba2c6b5605cde27ce), is anchored to Bitcoin blocks 948142, 948151, and 948161.

Release runbook · Timestamping ceremony


Use it from the command line

git clone https://github.com/openIE-dev/free-humanoid-corpus.git
cd free-humanoid-corpus

# Validate the corpus and your local edits
python3 tools/validate.py corpus.jsonl --strict

# Look up prior art for a patent-claim phrase
python3 tools/lookup.py "harmonic drive reducer with output torque sensor"

# Scaffold a new entry interactively (then submit as a PR)
python3 tools/new_entry.py

License

CC0 1.0. Public domain dedication. The corpus exists to be cited, copied, indexed, mirrored, embedded, ingested, and built upon. No usage license required.


Public domain (CC0 1.0). The corpus IS the prior art commons.

This site uses Just the Docs, a documentation theme for Jekyll.