Ragenta Docs
Models

Models

The workspace's default chat and embedding models, the catalogue table, tiers by plan, and the order in which a model is chosen.

Settings → Models. Viewing needs model.read; changing defaults needs model.manage (Owner/Admin).

Workspace defaults

FieldNotes
Chat modelUsed by Chat and agents when nothing else is chosen.
Embedding modelUsed when a knowledge base is created. Changing this affects knowledge bases created from now on — existing ones keep theirs.

Catalogue

Every model the deployment knows about: Model · Provider · Use · Tier · Context · Status.

  • Use: chat, embedding, rerank.
  • Tier: economy or premium. Free and Starter get economy only; Pro and up get premium too.
  • Status: available · no provider key (the deployment has not configured that provider) · not in your plan (tier outside the plan).

An alert No provider is configured at the top means no provider is live at all — ask the deployment administrator; API keys cannot be added at workspace level.

How a model is chosen for an answer

  1. The model pinned on the Project (if the conversation/run chose a project with a pin).
  2. The model picked on the Chat toolbar or in the agent configuration.
  3. The workspace's default Chat model.
  4. The deployment's default economy model.

Models cost differently — see Credits by model on Usage.

Webhooks — tell your systems when something happens

Register an endpoint, the available events, verifying the X-Ragenta-Signature header and reading the delivery log.

Credits

What a credit is, the two credit buckets (plan and top-up), the spending order, and what happens when you run out.

On this page

Workspace defaultsCatalogueHow a model is chosen for an answer