Skip to main content

Models

Route: /models · Settings section

What it is​

Where you register LLM providers and model configs agents can use (Anthropic, OpenAI, Ollama, …), including pricing metadata for Costs.

What it is for​

  • Register LLM providers and model configs agents can use
  • Usually done once — often already handled in the Welcome wizard
  • Adjust later for cost (Costs) or different agents

You only need this screen before an agent runs — not to install oc8 itself.

Where you are in the flow​

★ Models configured → assign on Agents → runs consume tokens → Costs

Without a model, agents cannot usefully run. This is early setup (also linked from the welcome / pilot checklist).

What you do here​

  1. Add a provider (API key via env or UI, depending on deployment).
  2. Declare / select models.
  3. On each Agent, pick the model config.
  4. Prefer cheaper models for pilots; tighten later.

Where work goes next​

GoalNext
Run an agentAgents
Local model without cloud keysOllama profile — Quickstart
Watch spendCosts