Quickstart
A first live result in under two minutes — Model Intelligence health check, model lookup, and where to go next.
Reach a first live result in under two minutes. No account, no API key: the Model Intelligence public API answers anonymously.
1. Check the service is up
curl -sS https://upcube.ai/api/mi/v1/healthExpected: HTTP 200 with ok: true plus seed counters. If you see anything else, stop here and read Troubleshooting — downstream steps depend on this one.
2. List models
curl -sS https://upcube.ai/api/mi/v1/modelsExpected: HTTP 200 with a models array and a total count. Each entry carries a slug you can look up next.
3. Look up one model
Pick a slug from step 2 and fetch its envelope:
curl -sS https://upcube.ai/api/mi/v1/models/<slug>Expected: HTTP 200 with the model's full envelope, or HTTP 404 {"error":"not found"} for an unknown slug. Removed models answer 410; quarantined models answer 500 with a reason. Full semantics: Model Intelligence API.
4. Open the apps
- Ethen Chat — sign in and start a conversation (Chat overview)
- Model Intelligence — browse the directory this API serves
- Studio — private alpha; enrollment required (Studio overview)