> ## Documentation Index
> Fetch the complete documentation index at: https://docs.llmrouter.app/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenClaw (CLI)

> Configure the OpenClaw terminal agent to use LLM Router for intelligent routing, context compression, and Zero Data Retention (ZDR).

***

> ## Documentation Index
>
> Fetch the complete documentation index at: [https://docs.llmrouter.app/llms.txt](https://docs.llmrouter.app/llms.txt)
> Use this file to discover all available pages before exploring further.

# OpenClaw & LLM Router

[OpenClaw](https://github.com/openclaw-ai/openclaw) is a powerful open-source AI coding agent designed for the terminal. It excels at autonomous code generation, project scaffolding, debugging, and multi-step development tasks.

Because autonomous agents like OpenClaw rapidly execute commands, read large files, and ingest massive terminal outputs, they are highly prone to **token bloat** and **accidental secret leaks**.

By connecting OpenClaw to **LLM Router**, you automatically apply **Middle-Out Context Compression** to bloated terminal histories and ensure your local `.env` secrets are **Redacted** before they reach third-party AI providers.

***

## Step 1: Installing OpenClaw

You can install OpenClaw using one of the following methods:

<CodeGroup>
  ```bash npm (Recommended) theme={null}
  npm install -g openclaw
  ```

  ```bash Homebrew (macOS) theme={null}
  brew install openclaw
  ```

  ```bash Source theme={null}
  git clone https://github.com/openclaw-ai/openclaw.git
  cd openclaw
  npm install
  npm link
  ```
</CodeGroup>

***

## Step 2: Configuring LLM Router

OpenClaw natively supports custom OpenAI-compatible providers, making it incredibly simple to connect to LLM Router.

<Steps>
  <Step title="Get Your API Key">
    Log into the [LLM Router Dashboard](https://llmrouter.app/dashboard) and generate a new API key (e.g., `sk-router-...`).
  </Step>

  <Step title="Edit the OpenClaw Config">
    OpenClaw uses a local JSON file for configuration. Open your terminal and run
    the configuration command: `bash openclaw config ` Alternatively, you can
    manually edit the file located at `~/.openclaw/config.json`.
  </Step>

  <Step title="Add LLM Router Settings">
    Update the configuration file to point the `baseUrl` to LLM Router. Make sure the `provider` is set to `"openai"`.

    ```json ~/.openclaw/config.json theme={null}
    {
      "provider": "openai",
      "baseUrl": "https://api.llmrouter.app/v1",
      "apiKey": "sk-router-your-api-key-here",

      // You can use any model supported by LLM Router
      "model": "anthropic/claude-3-5-sonnet",
      "temperature": 0.7
    }
    ```
  </Step>
</Steps>

<Note>
  **Model Selection:** Because LLM Router acts as a universal gateway, you must
  prefix your chosen model with the provider slug (e.g.,
  `anthropic/claude-3-5-sonnet`, `deepseek/deepseek-chat`, or `openai/gpt-4o`).
</Note>

***

## Step 3: Start Using OpenClaw

With the configuration saved, you can now launch OpenClaw. All traffic is securely routed through LLM Router.

Run OpenClaw in interactive mode:

```bash theme={null}
openclaw
```

Or pass a specific task directly:

```bash theme={null}
openclaw "Implement user authentication with JWT and refresh tokens"
```

### The LLM Router Advantage

While OpenClaw builds your features, LLM Router is working in the background:

* **Token Savings:** When OpenClaw reads a 2,000-line file, LLM Router dynamically compresses the irrelevant middle sections if the context shifts, saving you up to 80% on API costs.
* **Security (ZDR & Redaction):** If OpenClaw accidentally reads an AWS Key from your local workspace, LLM Router's PII engine will redact it into `[TOKEN_REDACTED]` before sending the code to Anthropic or OpenAI.
* **Reliability:** If the primary AI provider experiences an outage, LLM Router will automatically fall back to a secondary provider without breaking OpenClaw's autonomous loop.