Enjoy waived service fees on top-ups, plus exclusive referral gifts
Open SourceMIT License
OpenClaw

OpenClaw

The AI that actually does things.

An open-source local AI automation platform that runs agents through your existing 20+ messaging channels to take real actions.

Supported channels

WhatsAppTelegramSlackDiscordSignaliMessageTeamsGoogle ChatMatrixEmailVoiceSMS+10 more

Capabilities

Not just chat β€” it takes action

20+ Messaging Channels

Run your AI agent through WhatsApp, Telegram, Slack, Discord, iMessage, and more β€” from a single local gateway.

Browser & System Automation

Browse the web, fill forms, read and write files, and execute shell commands β€” with configurable sandboxing.

Voice & Device Integration

Wake word detection, continuous voice interaction, camera, screen capture, and location β€” across macOS, iOS, and Android.

Persistent Memory & Context

24/7 context maintenance, proactive cron jobs, reminders, and memory of your preferences across all sessions.

Multi-Agent Architecture

Run parallel AI instances for different purposes or users via a WebSocket gateway control plane with tool streaming.

Local-First, Zero Cloud Dependency

Runs entirely on your own infrastructure. Your data never leaves your machine unless you choose to connect external APIs.

Better together

One API key, 500+ models

OpenClaw natively supports any OpenAI-compatible API. Point it at OneRouter's endpoint, and your agent can seamlessly switch between GPT, Claude, Gemini, DeepSeek, and hundreds more β€” with governance and unified billing applied automatically.

500+ models via one endpoint
Unified governance & cost controls
Team-level access policies
Get a OneRouter API Key

Install OpenClaw

bash
npm install -g openclaw@latest
bash
openclaw onboard --install-daemon

Configure OneRouter as your model provider

yaml
# ~/.openclaw/config.yaml
model_provider:
  base_url: https://onerouter.com/v1
  api_key: YOUR_ONEROUTER_KEY
  default_model: gpt-4o    # or claude-opus-4.5, gemini-3-pro...

Once configured, all OpenClaw agent requests route through OneRouter with model policies and usage tracking applied automatically.

FAQ

Frequently Asked Questions

OpenClaw supports any OpenAI-compatible API, including Claude (Anthropic), OpenAI GPT, and local models. By pointing it at OneRouter's endpoint, you get access to 500+ models through a single API key.
Yes. OpenClaw is MIT licensed and completely free. You only pay for the API usage billed by your model provider β€” in this case, through your OneRouter workspace.
OpenClaw is a local automation platform, not a chat interface. It acts as an agent that can actually take actions β€” sending messages, browsing the web, running scripts β€” across your existing communication channels.
OneRouter gives OpenClaw access to 500+ models through one endpoint, with governance, cost controls, and team-level access policies β€” instead of managing API keys for every provider separately.

Deploy your local AI agent

Install OpenClaw, connect OneRouter, and let AI actually work inside your existing workflows.