NewAPI

Cherry Studio has built in support for the New API provider. A single New API address can connect to OpenAI Chat, OpenAI Responses, Anthropic Messages, and Gemini protocols at the same time, without needing to add a separate provider for each protocol.

Configuration steps

1. Create and copy a token in the New API admin panel;
2. Open Cherry Studio [Settings] → [Model Services] → [New API];
3. Fill in [API Key];
4. In [API Address], enter the root address of the New API, for example https://api.example.com;
5. Obtain or manually add the model, and after completing [Test], turn on the enable switch in the upper right corner.

Cherry Studio will automatically use the corresponding version path according to the selected protocol, for example /v1 or /v1beta. Usually you only need to enter the root address once; do not paste the full endpoint path such as /chat/completions into it.

If you have previously used the beta version and kept old values such as localhost:3000 in other protocol addresses, go to the New API settings, clear the old address for the non-default protocol once, and then test again. Otherwise, some models may still request the old address.

What should I do if the proxy station uses a path without a version?

Only when the proxy station explicitly requires a path without /v1 or /v1beta should you append #to the end of the corresponding API address.# This means version paths will no longer be appended automatically; do not use it for a normal New API deployment.

The address must keep the correct http:// or https://. When using an IP and port, you can enter a root address such as http://127.0.0.1:3000 .

---

Get help and submit feedback

If you encounter any questions, bugs, or feature improvement suggestions during configuration or use, please refer to the official channels provided in Feedback and Suggestions .