Skip to content

Models

Purpose: Select and configure the models you want to use with your preferred parameters.

Models screen — Models tab with a Provider selector on the left, a list of provider model ids discovered in the middle, and a right‑hand form showing “Model Name”, “Description”, “Advanced Parameters” (Temperature, Max Tokens, Top‑P/Top‑K, Seed, Stop), and a Save button.

Typical flow

  1. Select Provider → select Provider Model (vendor model id).
  2. Model Name is automatically set to the default name from the provider. You can change this in advanced options.
  3. Optionally set a different Model Name (your friendly id used everywhere, e.g., gpt-4.1-mini-factual), Model Description and Advanced Parameters (Temperature, Max Tokens, Top‑P/Top‑K, Seed, Stop).
  4. Click Save.

Advanced parameters

  • Keep temperature low for factual agents; raise for creative tasks.
  • Only include parameters you need; otherwise defaults are applied by the provider.

API parity

  • Discover vendor models: POST /v1/core/models
  • Add/save: POST /v1/core/add/model
  • Delete: POST /v1/core/delete/model
  • List (all regular models plus chat-eligible agents): GET /v1/models
    See Models

Gotchas

  • If a model fails at runtime, confirm the provider key scope and that your account has access to that model tier.