OpenClaw

Overview

Maintainer	OpenClaw Foundation (non-profit) — originally by Peter Steinberger
Type	Local daemon + messaging gateway
Pricing	Free — open source, bring your own model
Platforms	macOS, Linux, Windows
Protocol	MCP, agent skills
Website	openclaw.ai

What It Does

OpenClaw is a personal AI assistant that runs on your own machine and reaches you wherever you already are. It controls a browser, reads and writes files, runs shell commands, and keeps memory across conversations.

The scope is wider than a coding agent. It organizes an inbox, sends email, manages a calendar, and checks you in for flights, and it happens to be able to write code as well. If you want a harness built only for repositories, a terminal coding agent is the closer fit; if you want one assistant that handles the rest of the job around the code, this is the shape of it.

Install

curl -fsSL https://openclaw.ai/install.sh | bash

Or through npm:

npm i -g openclaw

A source checkout with pnpm works too.

Key Strengths

• 29 messaging channels: WhatsApp, Telegram, Discord, Slack, iMessage, Signal, and more — the agent is not stuck in a terminal
• Runs locally: your machine, your files, your keys
• Writes its own skills: skills can be created through ordinary conversation rather than authored by hand
• Persistent memory: context carries across conversations instead of resetting each session
• Real computer access: browser control, filesystem, and shell
• Community maintained: a foundation rather than a single vendor, which matters for something holding this much access

Trade-offs

• Broad rather than specialized — a dedicated coding agent will go deeper on large refactors
• An agent with your inbox, calendar, and shell is a large amount of trust in one process, so review what you grant it
• Self-hosted, so the deployment and updates are yours

Where to Find Skills

OpenClaw can write its own skills, and it can also use ones other people wrote. The largest open catalogue is atskills.one, with 60,000+ skills — see @skills for how referencing and saving work.

Best For

• One assistant across chat, browser, and shell rather than a tool per surface
• Personal and operational automation that surrounds the coding work
• Local-first setups where the agent should not be someone else’s cloud

---