ModelTrust
Menu

Verification console

Check an AI endpoint.

Submit an OpenAI-compatible endpoint. Level 1 sends three minimal, bounded requests and reports only directly observed operational evidence.

What you can test now

Level 1 checks an OpenAI-compatible chat endpoint with three bounded requests.

What it does not verify

It does not prove model identity, benchmark intelligence, safety, or future availability.

How your key is handled

The key is used server-side for this run only. It is never stored, logged, returned, or sent to analytics.

Verification mode

A label for this report; it does not need to match a legal entity.

Public HTTPS origin and optional API prefix, for example /v1. No path to /chat/completions.

The exact model value your integration sends to the endpoint.

Used once on the server and never persisted.

Secret handling boundary

Your key is sent only to the ModelTrust server for this verification. It is never stored, logged, included in analytics, or returned in the report.

This run sends 3 Level 1 requests. Each is limited to 10 seconds and the complete run to 35 seconds.