678APIdocs
Use 678API in Cherry Studio
Add 678API through Cherry Studio custom providers, then manage its models and conversations from the desktop app.
678APIOpenAIhttps://www.678api.com/v1Key created in the consoleExact model name from the consoleClick Settings in Cherry Studio and open Model Services.
Click Add, use 678API as the name, select OpenAI as the provider type, and save.
Select the new provider and paste an API key created in the 678API console.
Use https://www.678api.com/v1. Do not append /chat/completions or another endpoint path.
Click Check beside the API key. If it fails, inspect the URL, key, balance, and network.
Add the exact model IDs shown in the console, or fetch the model list when your Cherry Studio version supports it.
Enable the provider, choose a 678API model in Chat, and send a short test message.
Common issues
| Symptom | Fix |
|---|---|
| Key check fails | Use https://www.678api.com/v1, copy the key again, and confirm available credits. |
| Model list is empty | Add model IDs manually from the 678API console. |
| Model not found | Check case, hyphens, version suffix, and account group permissions. |
| No streaming output | Enable Stream in model settings and retry with an updated Cherry Studio version. |
You do not need to sign out of Cherry Studio or another provider. Each chat uses its selected provider and model; select the previous provider to switch back.