Configure an AI provider

Open AI Providers

Open Settings → AI Providers. Select a built-in provider or choose Add custom, then enter the connection values supplied by that provider.

After saving, use model sync to retrieve the available model list and choose the default model used by new Agents and chats.

Built-in defaults

• OpenAI base URL: https://api.openai.com/v1; wire: responses
• Anthropic base URL: https://api.anthropic.com; wire: anthropic_messages
• DeepSeek base URL: https://api.deepseek.com/v1; wire: responses
• Kimi base URL: https://api.moonshot.cn/v1; wire: responses
• Custom providers can set Base URL, models endpoint, auth header/scheme, and one of the two supported wire APIs

Wire compatibility

Kition 0.1.41 supports responses and anthropic_messages. Legacy Chat Completions is not a supported fallback. A service exposing only /chat/completions is not sufficient unless it also provides a compatible supported wire.

Credentials and model sync

• Use a project-scoped or limited provider key when the provider offers one
• Save the provider before syncing models
• Copy the model identifier returned by the provider rather than guessing from a display name
• Choose a default model after sync; switch per chat when a task needs another model
• Disconnect to disable a provider, then revoke or rotate its remote key; 0.1.41 does not reliably delete the saved local secret through this UI action

Common connection failures

• 401 or 403 — invalid, expired, incomplete, or unauthorized credential
• 404 — incorrect Base URL, models path, wire, or model identifier
• 429 — provider request/token quota reached
• No usable models — the models endpoint returned no compatible text-generation models
• Connection error — proxy, certificate, DNS, or local service availability problem