Connecting Your Own AI Model

Oyren AI never locks you into one model provider. This guide shows how to bring your own API key — or point at a local model — so chat and agent runs use exactly the provider you choose.

1. Open model settings

From your workspace, go to Settings → Models. This is where every provider key you add is scoped — keys are never shared across workspaces you don't control.

2. Add a provider key

Paste in an API key from any supported provider:

• OpenAI
• Anthropic
• Google
• Mistral
• Cohere
• A local model endpoint (e.g. Ollama)

Settings -> Models -> Add provider
  Provider: Anthropic
  API key:  sk-ant-••••••••••••

Your key is encrypted at rest and used only for requests you initiate — Oyren AI doesn't proxy your usage through a shared pool.

3. Pick a default model per workspace

Different workspaces can use different models. A workspace for quick drafts might default to a fast, cheap model, while one for legal review defaults to a stronger reasoning model — set this per workspace from the same Models screen.

4. Switch models mid-conversation

You're not locked into the default once a conversation starts — switch models from the chat panel at any point, and the new model picks up the full conversation history and any grounded citations already in play.

What's next

Pair this with Getting Started: From PDF to Your First Chat if you haven't uploaded a document yet, or explore agent tool permissions to control what an agent can do with your chosen model.