Quick answer: In Cherry Studio go to Settings → Model Providers → Add, choose OpenAI as the type, set the API host to https://321ai.cc/v1, paste your 321AI key, then click "Fetch/Check" to pull the model list — if your version cannot fetch, add model IDs manually.
Audience: desktop users of Cherry Studio who want one endpoint for GPT / Claude / Gemini / Qwen models.
| Setting | Value | Note |
|---|---|---|
Provider type | OpenAI | Cherry Studio built-in type |
API Host | https://321ai.cc/v1 | include /v1 |
API Key | your 321AI key | from the 321AI console |
Models | fetch list or manual IDs | exact names from the 321AI catalog |
Confirm the host includes /v1 and the key is valid; some Cherry Studio versions require saving the provider before fetching.
Reduce the max-tokens setting or try another model; check the console usage log to confirm the request completed.
Related: 中文版 Cherry Studio 教程 · monthly cost calculator