Certify a model, agent, or autonomous bot

Certification is a contract: it names who owns and built the model, agent, or autonomous bot, where it originates, what it is capable of, its scale, whether it fingerprints its output, whether it makes agentic decisions, and what unintended consequences have been reported. Fill in the form to download a certification file, or upload one to render it below.

To list your certification in the public registry, download it below, add the JSON file to src/data/certifications/ in this repository, and open a pull request. Anyone can review it before it goes live.

What are you certifying?
Capabilities (modalities)
Fingerprinting
Hazard categories to disclose

HAZMAT-style placards. Pick from the list or add your own.

Data-handling & compliance standards

Sensitive data classes and regulations this model or agent is certified to handle safely, e.g. PII, HIPAA, GDPR, SECRET.

Approved operating regions

Where this model or agent is cleared to operate.

Compliance standards and operating regions are self-disclosed by the owner, not independently audited by this project. Treat them as a claim to verify, the same as any other field on the certification.

NIST AI RMF functions addressed

Which NIST AI Risk Management Framework functions this certification maps to.

Decision categorization & sub-categorization (yearly disclosure)
Incident references (public URLs)

Links to logged incidents behind the consequences above, e.g. an AI Incident Database entry.

Signature (optional, for provenance)

Paste a detached signature over this certification so others can verify who issued it. Sign the JSON with this field removed.

Upload a certification

Have a certification file already? Upload it to render it below and load its values into the form above, so you can edit it and download a new certification. Older files missing newer fields still load; the missing fields start empty.

A global, plain-language guide to AI development. Not legal advice, not a government filing.