Connect Yunhai from a client that supports a custom model API.
Connection settings
| Provider | OpenAI Compatible / Custom API |
|---|---|
| Base URL | http://localhost:3187/v1 |
| API key | A project key created in the Yunhai console |
| Model | A model ID available to your key group |
Configure the client
Open the client model or provider settings, choose a compatible protocol and enter these values. Setting names vary by client version.
Verify the connection
Start with a short text request before enabling tools, images or long-context tasks. If a full Chat Completions endpoint is required, use:
text
http://localhost:3187/v1/chat/completionsOpenClaw runtime environments, messaging plugins and cloud instances require separate deployment. A Yunhai model key grants model API access only.
Troubleshooting
First verify the key and model using the quick-start request, then investigate client settings.