Using LiteLLM With Openclaw

Standardizing multiple AI providers through a single proxy simplifies your stack. Here’s what you need to know:

1. LiteLLM matters because it unifies dozens of providers under one OpenAI-compatible API, cutting integration time significantly.
2. Developers often struggle with proxy setup and configuring the right API format for OpenClaw compatibility.
3. You will learn how to start the LiteLLM proxy and connect it to OpenClaw using both interactive and manual methods.

LiteLLM acts as a universal translator between your application and various AI providers. It exposes an OpenAI-compatible endpoint that OpenClaw can consume seamlessly.

Start by installing and running the LiteLLM proxy locally:

pip install 'litellm[proxy]'
litellm --model claude-opus-4-6

The proxy runs on http://localhost:4000 by default. Now configure OpenClaw using the interactive wizard:

openclaw onboard --auth-choice litellm-api-key

For manual configuration, define your models in openclaw.json:

{
"models": {
"providers": {
"litellm": {
"baseUrl": "http://localhost:4000",
"apiKey": "${LITELLM_API_KEY}",
"api": "openai-completions",
"models": [
{
"id": "claude-opus-4-6",
"name": "Claude Opus 4.6",
"reasoning": true,
"input": ["text", "image"],
"contextWindow": 200000,
"maxTokens": 64000
}
]
}
}
},
"agents": {
"defaults": {
"model": { "primary": "litellm/claude-opus-4-6" }
}
}
}

Restart OpenClaw and your proxy becomes the single entry point for all model requests.