World Archive

Overlap · catalog comparison

Someone is selling you data. Is it already free?

Compare a dataset offer against fingerprints of openly available robotics data. Re-encodes, trims, splices, mirrored copies and slowed-down footage are all caught — a checksum catches none of them.

Drop a manifest

The .ovlm file overlap export wrote. If you exported with --split-gb, drop every part and its parts.json.

No manifest yet? Three commands

  1. 01

    Install the tool

    $ pip install git+https://github.com/World-Archive/overlap
  2. 02

    Fingerprint the footage you were offered

    $ overlap --index ./offer.ovl index /path/to/offer
  3. 03

    Export a manifest

    $ overlap --index ./offer.ovl export -o offer.ovlm --stride 1

Full instructions, including split exports.

What leaves your machine

Only fingerprints. A manifest holds 256-bit perceptual hashes and file metadata — never frames, never pixels. It is still a confidential business document: it names your files, and someone holding it can confirm whether footage they already have matches yours.

Your manifest and its result are stored so the page keeps working and can be re-run against datasets added later. The result lives at an unguessable link that search engines are told to skip — anyone with the link can read it, so share it deliberately.