The base URL of your OpenAI-compatible gateway, for example:

• https://litellm.example.com for a LiteLLM proxy
• https://openwebui.example.com/api/v1 for an OpenWebUI backend
• http://localhost:8000/v1 for a self-hosted proxy

/chat/completions is appended automatically. The URL must use http or https.

HTTP redirects (e.g. http to https upgrades behind load balancers) are followed automatically.