Configure LLM Keys
Add an AI provider API key so your agent can think and respond.
Your agent needs an API key from an AI provider to think and respond. This page explains how to add one.
Why you need an LLM key
AI agents don't have built-in intelligence — they call external AI models (like GPT-4, Claude, or Gemini) to generate responses. To do this, they need an API key from the provider.
You choose and pay for your own AI provider. This means you control your costs, your model choice, and your data.
Supported providers
RunClaw agents support a wide range of AI providers:
| Provider | Models | Get API Key |
|---|---|---|
| OpenAI | GPT-4o, GPT-4, GPT-3.5 | platform.openai.com |
| Anthropic | Claude 4, Claude 3.5 | console.anthropic.com |
| Gemini 2.5, Gemini 2.0 | aistudio.google.com | |
| OpenRouter | 100+ models from multiple providers | openrouter.ai |
| DeepSeek | DeepSeek-V3, DeepSeek-R1 | platform.deepseek.com |
| Mistral | Mistral Large, Codestral | console.mistral.ai |
| xAI | Grok | console.x.ai |
OpenRouter is a great choice if you want to try multiple models without managing separate API keys. One key gives you access to models from OpenAI, Anthropic, Google, Meta, and more.
How to add your key
For OpenClaw agents
- Go to your RunClaw dashboard
- Click on your instance
- Go to Settings > Configuration
- Find the LLM provider section
- Select your provider and paste your API key
- Click Save
The key is written directly to your VPS configuration file. It never passes through RunClaw's servers — the dashboard communicates with your VPS through an encrypted sidecar connection.
For Agent Zero agents
- Go to your RunClaw dashboard
- Click on your instance
- Go to Settings > Agent Zero Config
- Set your chat model provider and name (e.g.,
openai/gpt-4o) - Paste your API key
- Click Save
Agent Zero has four model slots: Chat, Utility, Browser, and Embedding. The utility and browser models default to your chat model settings if not set separately.
Where your keys are stored
Your LLM API keys are stored on your VPS only:
- OpenClaw: Written to
openclaw.jsonon the host filesystem - Agent Zero: Written to
settings.jsonand.envinside the Docker container
They are never sent to or stored on RunClaw's servers. The sidecar connection that carries configuration changes uses end-to-end encryption between your browser and your VPS.
You can verify this by SSHing into your VPS and checking the config files directly. You have full root access.
Troubleshooting
"Invalid API key" errors: Double-check that you copied the full key. Some providers generate keys with a prefix (e.g., sk-... for OpenAI).
Model not available: Make sure your API plan supports the model you selected. Some models require a paid tier at the provider.
Agent not responding after adding key: The agent container may need a restart. Go to your dashboard and click Restart on the instance.
You're all set. Your agent is deployed, connected, and ready to work.
Back to Getting Started