OpenClaw is an open-source personal AI assistant that runs as a self-hosted gateway across messaging apps, devices, agents, tools, and model providers.

The OpenClaw project is free and open source, but users may still pay for model APIs, hosting, local hardware, or third-party services connected to the agent.

Who should use OpenClaw?

OpenClaw is best for developers, founders, and technical power users who want a self-hosted assistant they can reach from chat apps and control directly.

Which model providers does OpenClaw support?

OpenClaw is model-agnostic and can be configured with providers such as OpenAI, Anthropic, OpenRouter, MiniMax, and local model services depending on setup.

Can OpenClaw run on my own machine?

Yes. OpenClaw is designed to run on your own hardware or server, with a gateway, dashboard, workspace, sessions, and channel configuration under your control.

Is OpenClaw safe for sensitive workflows?

It can be used carefully, but users should review permissions, provider keys, channel access, allowlists, and connected tools before giving an agent sensitive access.

OpenClaw

Name: OpenClaw
Brand: OpenClaw

Open-source personal AI assistant that runs across chat apps and devices.

Visit Website

What is OpenClaw?

TL;DR

OpenClaw is an open-source AI agent platform for running assistants that can connect to tools, messaging, files, and local or hosted workflows to perform practical tasks. It is aimed at teams and builders who want practical, repeatable workflows with AI.

Multi-channel gateway for messaging apps
Self-hosted agent runtime with local control
Model-agnostic provider routing and failover

Explore with AI

About

OpenClaw is an open-source personal AI assistant that runs as a self-hosted gateway across the chat apps and devices people already use. Instead of forcing every task into a single web app, it lets users message an always-available agent from places like Slack, Telegram, WhatsApp, Discord, Signal, iMessage, Google Chat, Microsoft Teams, WebChat, and other channel plugins. The product is built for developers and power users who want more control than a hosted assistant normally gives. OpenClaw runs on the user’s own machine or server, keeps gateway configuration and workspace state under their control, and supports model providers such as OpenAI, Anthropic, OpenRouter, MiniMax, and local model services through configurable routing. OpenClaw is strongest when the assistant needs to keep working beyond one chat session. It supports stateful sessions, memory, tools, skills, cron-style reminders, webhooks, browser and file workflows, mobile nodes, and a browser dashboard for controlling the gateway. That makes it useful for personal automation, coding coordination, inbox and calendar support, research, follow-ups, and cross-device task execution. The tradeoff is setup and responsibility. Teams must understand provider keys, channel permissions, allowlists, local security, and what the agent can access before using it for sensitive work. For technical users who want an inspectable, model-agnostic, self-hosted AI assistant, OpenClaw is one of the more serious open-source agent frameworks to evaluate.

Use Cases

Personal AI Agents

Run assistants that can connect to tools, files, and services to handle practical tasks.

Workflow Automation

Automate repeatable work by giving agents access to relevant tools and actions.

Developer Agent Work

Build and customize agent workflows for coding, research, and operational tasks.

Local AI Automation

Use an open-source approach when you want greater control over agent deployment and integrations.

Key Features

Multi-channel gateway for messaging apps
Self-hosted agent runtime with local control
Model-agnostic provider routing and failover
Skills, memory, cron, tools, and webhooks

Featured Tools

Wispr Flow

Voice-First Dictation

We may earn commissions from links to support our work. Learn more.