Models, providers and routing

Managed model aliases

Publish a stable name like chat-default, then repoint it to a better or cheaper model without touching a single app.

What it is

Publish a stable name such as chat-default or best-coder, point it at any catalog model, and repoint it later. Callers keep sending the same name.

Why you want it

Model upgrades stop being client changes. One repoint moves every app and person to the new model on the next request.

Three apps all call the stable name chat-default. Over time the AI team repoints chat-default from Model A to the cheaper Model B and then the faster Model C, while a preview Model D stays disabled. Support bot Code review Docs search chat-default the name apps use Model Athe old default Model Bhalf the price Model Cfaster, smarter Model Dpreview · off The apps never change. Your AI team decides what chat-default means.

How it works

  • Aliases are granted independently and never chain
  • /v1/models shows what each alias currently resolves to
  • Reports record the model that actually ran plus the alias that was requested
  • A broken alias returns policy.managed_model_unavailable instead of 'model not found'
  • Every create, repoint and delete is audit-logged
Callers keep one name; the response says what ran
POST /v1/chat/completions
{ "model": "chat-default", … }

HTTP/1.1 200 OK
X-Janus-Model: janus-sol-5
X-Janus-Fallback-Reason:            # set only when a fallback served the request

In the product

Janus Managed models screen with demo data
Managed models

Stable names, what each resolves to right now, its fallback, and 30 days of spend, requests and callers.

Actual Janus interface · Demo identities and synthetic usage

See it on your own network.

The Community edition is free for up to 25 people. The 30-day Business trial unlocks every Business feature.