Your assistant needs a language model behind it. OpenClaw supports any provider that exposes an OpenAI-compatible chat completions endpoint. So which do you pick?
The 3 most common cloud choices are:
1. Anthropic (Claude). Strong at following instructions and refusing unsafe requests. Set your provider to anthropic and supply a Claude API key.
2. OpenAI (GPT). Broad ecosystem, function-calling support. Use provider openai with your OpenAI API key.
3. DeepSeek. Cost-effective for high-volume use. Works through OpenRouter or a direct DeepSeek API key.
You configure the provider and model in ~/.openclaw/model.json. I'll explain the exact format in the Model Refs unit.