.svg)
OpenClaw is a self-hosted personal AI assistant that runs on macOS, Windows, and Linux, connecting to the messaging channels users already rely on — WhatsApp, Telegram, Discord, Slack, Signal, iMessage, Google Chat, Microsoft Teams, BlueBubbles, Matrix, Zalo, and WebChat. Installed via a single command or cloned from source, it operates as a local-first Gateway that manages sessions, channels, tools, and events entirely from the user's own machine.
The assistant supports persistent memory, meaning context accumulates across conversations and carries over between sessions. It connects to Anthropic Claude and OpenAI models, with the recommended configuration being Anthropic Pro or Max with Opus 4.5. Users can also route through local models. Browser control via an openclaw-managed Chrome or Chromium instance allows the assistant to browse, fill forms, and extract data from any website.
Full system access enables the assistant to read and write files, run shell commands, and execute scripts, with a sandboxed mode available for group channels and non-main sessions. A companion macOS menu bar app provides local control alongside optional iOS and Android node apps that support Voice Wake, Talk Mode, camera, and screen recording. The skills platform — backed by ClawHub — allows users to install community-built skills or build their own, and the assistant can author and hot-reload new skills through conversation.
Automation capabilities include cron jobs, webhooks, Gmail Pub/Sub triggers, and a multi-agent routing system that can send messages across sessions. Security defaults include DM pairing, allowlists, and per-session elevated access controls. OpenClaw is open-source under the MIT license with an active contributor community.
- Control Gmail, calendar, and browser from WhatsApp or Telegram via a local AI assistant
- Run autonomous Claude Code or Codex loops remotely and receive progress updates in chat
- Set up cron jobs and webhook triggers so the assistant completes tasks on a schedule
- Manage files, run shell scripts, and execute system commands from any messaging app
- Extend assistant functionality by installing or building custom skills through conversation
- Route messages across multiple sessions for multi-agent task coordination
- Enable Voice Wake and Talk Mode on macOS and iOS for always-on speech interaction
- Use the browser control tool to fill forms, extract data, and automate web tasks
- Deploy the assistant on a remote Linux server and access it via Tailscale or SSH
- Configure DM pairing and allowlists to control who can interact with the assistant
