Providers
Switch the active provider and model
Understand the device-wide model selection that Kition 0.1.41 saves when you sync a provider or use the Agent model picker.
Updated
By Kition Docs
Verified against Kition 0.1.41 (d3e1b931) on 2026-08-29
The current selection model
Kition 0.1.41 saves one active provider plus one selected model for each connected provider in desktop application settings. The active selection is device-wide and is not stored as a separate default inside each workspace.
There is no three-level global/workspace/task default hierarchy in this release. Changing the Agent model picker updates the saved active provider and selected model.
Note
Because this setting lives outside the workspace, copying a workspace to another device does not copy its provider selection or credentials.
Activate a provider through model sync
- Open Settings → AI Providers and select a configured provider
- Choose Sync now
- On success, Kition enables that provider and makes it active
- If the previous selected model is missing, Kition selects the first returned model
- Review that automatic choice in the Agent model picker before starting work
Switch from the Agent
The Agent model picker lists discovered models from enabled providers. Selecting an entry saves both its provider and model, so the next Agent request and other AI surfaces that use the desktop model settings can follow the new active selection.
Changing the selection does not cancel or rewrite a request that is already running. Wait for the current turn to finish or stop it before relying on the new model.
What switching does not do
- It does not delete credentials for the previous provider
- It does not move provider billing to Kition Cloud or between external providers
- It does not guarantee that every returned model supports Agent tools, images, reasoning controls, or hosted web search
- It does not create a workspace-specific override
If a model is missing or unusable
- Run Sync now again after the provider grants model access
- Confirm the provider remains enabled and its credential has not expired
- Choose a text-generation model rather than an embedding, image, audio, or moderation model
- Check that the model supports the configured
responsesoranthropic_messageswire - Disconnect and reconnect only after preserving any provider-side information you need for diagnosis
Implementation sources
Related pages
Configure an AI provider
Use Kition Cloud or connect OpenAI, Anthropic, DeepSeek, Kimi, or a supported custom endpoint.
Set up OpenAI
Connect an OpenAI project key, sync the models available to that project, and use the Responses wire in Kition.
Set up Anthropic
Connect an Anthropic API key, sync the models available to its workspace, and use the Messages wire in Kition.
