HQ is model-agnostic. There is no required or default provider — agents work with whatever you connect. OpenClaw supports 25+ providers including OpenAI, Anthropic, Gemini, OpenRouter, Ollama, LM Studio, and other OpenAI-compatible servers.Documentation Index
Fetch the complete documentation index at: https://docs.yourhq.ai/llms.txt
Use this file to discover all available pages before exploring further.
Adding a provider
Pick a provider
Choose from the provider catalog. Each provider shows its auth method and what you need.
Complete the auth flow
The method depends on the provider:
| Auth type | How it works |
|---|---|
| API key | Paste your key directly — connected immediately |
| OAuth / device code | Follow the login flow in your browser, then complete in the UI |
| CLI reuse | Reuse an existing auth session from a local CLI tool |
| Local URL | Point to a local server (Ollama, LM Studio, vLLM) |
Per-agent model overrides
Different agents can use different providers. For example:- A coding agent on Claude
- A research agent on GPT
- A local agent on Ollama