Cherry Studio
Connect Cherry Studio as a custom OpenAI-compatible model service.
Use the Cherry Studio documentation for current installation and UI details.
Configure
Install and open Cherry Studio.
Open Settings → Model Services → Add.
Set the provider name to AI Gateway, provider type to OpenAI, enter your platform API key, and set the API address to http://127.0.0.1:11113/v1.
Use Fetch models or manually add the complete <FULL_MODEL_ID> copied from the Models page.
Enable the provider, use Check, then return to chat, select the model, and send a message.
The API address must not include /chat/completions. The model list comes from /v1/models; always use the complete public model ID.
Limitations
This guide covers conversation and standard OpenAI-compatible capabilities. Whether Cherry Studio exposes separate image, video, embedding, or other endpoints depends on the current version of its UI. Do not infer capabilities from a model merely being visible; check the Models page.
Troubleshooting
- 401: Check the platform API key.
- 404: Check that the API address uses the
/v1base and not an operation path. - Model ID error: Use the complete model ID from
/models. - Vision or tools fail: Confirm that the selected model advertises the required vision or tool capability.
Help us improve this page
Found something unclear, outdated, or incorrect?
Last updated on