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

# OpenClaw

> The hottest open-source command-line AI coding agent. Complete complex programming tasks autonomously via the OpenAI-compatible interface.

[OpenClaw](https://github.com/OpenClaw/OpenClaw) is an extremely popular open-source command-line AI coding agent. It aims to provide capabilities similar to Claude Code while remaining fully open-source and supporting custom API access.

## Installation

Make sure you have Python 3.10+ installed:

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

## Configuration & Launch

OpenClaw can be configured via environment variables or an interactive setup.

### Option 1: Environment Variables (Recommended)

<Accordion title="Mac / Linux (zsh)">
  ```zsh theme={null}
  echo 'export OPENAI_API_KEY="sk-your-api-key"' >> ~/.zshrc
  echo 'export OPENAI_API_BASE="https://api.ephone.ai/v1"' >> ~/.zshrc
  source ~/.zshrc
  ```
</Accordion>

<Accordion title="Windows (PowerShell)">
  ```powershell theme={null}
  [Environment]::SetEnvironmentVariable("OPENAI_API_KEY", "sk-your-api-key", "User")
  [Environment]::SetEnvironmentVariable("OPENAI_API_BASE", "https://api.ephone.ai/v1", "User")
  ```
</Accordion>

### Option 2: Interactive Setup

Run `openclaw` directly, and the program will guide you through the API configuration:

1. **API Key**: Enter your ePhone AI API Key
2. **Base URL**: Enter `https://api.ephone.ai/v1`
3. **Model**: Recommended `gpt-4o` or `claude-opus-4-5-20251101`

## Key Features

| Feature                 | Description                                                                                   |
| ----------------------- | --------------------------------------------------------------------------------------------- |
| **Autonomous Coding**   | Describe a task in natural language, and OpenClaw will read code, modify files, and run tests |
| **Multi-Model Support** | Perfectly compatible with OpenAI and Anthropic protocol models                                |
| **Security Sandbox**    | Requests user confirmation before executing potentially dangerous commands                    |
| **Context Awareness**   | Automatically analyzes project structure to locate relevant code                              |

## Common Commands

Inside the OpenClaw interactive session:

* `help`: Show all available commands
* `clear`: Clear current conversation history
* `exit`: Exit the program

## Recommended Models

| Use Case                  | Recommended Model          |
| ------------------------- | -------------------------- |
| Complex logic development | `claude-opus-4-5-20251101` |
| General coding tasks      | `gpt-4o`                   |
| Fast fixes and queries    | `gpt-4o-mini`              |

<Note>
  As a powerful agent, OpenClaw performs significant context reading and model calls, leading to faster token consumption. We recommend using a Tier 2 account for a more stable experience.
</Note>