Submit

List a pack, or ask to be measured

Listing is free and public facts only. A verification run adds the one number nobody else publishes: how often your pack agrees with a real expert.

Option 1 · 3 minutes

The form

A GitHub issue with the public facts: name, vendor, platforms, profession, pricing, license, links, and the three AI Act answers. We review it and add the YAML.

Open the form

Option 2 · pull request

The YAML

Add a file under catalog/packs/ following the schema. CI validates the schema and the AI Act rules before anyone reads it.

Open the repository

Verification

What a run costs and what you get

Tick the verification box in the form. We reply with a free 10-card pilot from an expert in your profession: a public report, no number yet. If the pilot shows a gain, a full run on at least 30 cards gives you the Verified badge, an embeddable badge that links to the report, and a re-run whenever your named model changes.

Pricing for full runs is published on the report of the first measured pack. Until then, every run is a pilot.

We never list packs that perform an AI Act Art. 5 practice, that decide on people without a human, or that automate LinkedIn. Synthetic cases only: a golden set never contains a real person or client.