Models & Providers

webAgent talks to AI models through providers — by default via OpenRouter, which gives access to many models (from various makers) through one connection. You configure which model the agents use in App Config.

Key ideas:

• Pick your model. Set the provider, model name, and key. The same model also powers features like memory and wiki search (via its embeddings).
• Per-model roles (TEXT / IN / OUT). In the Models grid you mark which models handle text (the agent's brain — must support tools), see images (Image-in), and make images (Image-out). The app detects these when you save a model, and you can override them. Each job is routed to the right model; the app even cross-checks the model catalog so a mis-ticked text-only model can't be sent an image.
• Switching models per chat. A model can be chosen per conversation from the chat footer (or by the agent itself). See Choosing & Switching Models for the resolution order and the automatic fallback that keeps tool-less image models from breaking a turn.
• Run several providers at once (optional). You can configure multiple providers and have each message sent to all of them simultaneously, using the fastest complete answer. When an image is involved, only image-capable providers join the race.

Image generation uses whichever model you tick for Image-out; reading images uses an Image-in model. See Image Generation and Image Vision (Reading Images).

Related: Billing & Usage, Choosing & Switching Models.