OpenClaw
OpenClaw Foundation · Open-Source Personal AI Assistant
A free, open-source AI assistant that runs locally and works through the messaging apps you already use — Telegram, WhatsApp, Slack, Discord, Signal, iMessage, and 20+ more. Best for technical users who want a private, self-hosted agent that can browse the web, run commands, and automate personal tasks with the model of their choice.
Best for
Technical users who want a private, self-hosted AI assistant they fully control, running through their own messaging apps and model of choice
Not ideal for
Non-technical users or teams who want a managed, supported product with no setup or maintenance
Who it's for
Developers and technical power users who want a private, open-source personal assistant they self-host and connect to their own LLM
Capabilities
- Runs locally and operates through 29 messaging channels (Telegram, WhatsApp, Slack, Discord, Signal, iMessage, Microsoft Teams, Google Chat, Matrix, LINE, and more)
- Model-agnostic: 40+ bundled providers (Anthropic, OpenAI, Google, DeepSeek, xAI, Mistral, OpenRouter, and more), subscription sign-in, or local models
- Multi-agent routing that isolates channels and accounts into separate workspaces and sessions
- Extensible 'skills' system using SKILL.md files for custom tools and instructions
- Built-in tools: web browser, canvas, cron scheduling, and chat-platform actions
- Local storage of configuration and history for persistent, private behavior
- Optional voice wake words and talk mode on supported platforms
- Desktop apps for macOS, Windows, and Linux, plus iOS and Android nodes
- Self-hosted Gateway with WebSocket/RPC and inbound webhooks; ClawHub skills registry with a REST API
Limitations
- You must supply and pay for your own model access (API key, subscription, or local hardware) — there is no bundled model
- Self-hosted setup and maintenance still require technical comfort, despite the one-line installer and desktop apps
- Capability and reliability depend heavily on the external model you connect
- Autonomous actions (running commands, browsing) need human oversight to stay safe
- No hosted tier or paid support; the project is maintained by the non-profit OpenClaw Foundation and community
Use cases
- Running a private personal assistant inside your own chat apps
- Automating recurring personal tasks on a schedule via cron
- Drafting messages and replies across multiple messaging accounts
- Web browsing and lookups initiated from a chat conversation
- Building custom skills to extend the assistant for your own workflows
Our take
OpenClaw is best understood as the self-hosted, privacy-first answer to hosted assistants like Manus and Lindy. Instead of a polished SaaS you log into, it lives in your own messaging apps and runs on your own model key, so you trade turnkey convenience for control over your data and costs. That makes it a strong fit for developers and tinkerers, and a poor fit for anyone who wants something that just works without setup. Its momentum in 2026 — roughly 390,000 GitHub stars and a non-profit foundation behind it — reflects how much demand there is for an assistant you fully own.
Who should use it
Developers and technical power users who want a private, self-hosted assistant they can run in their own chat apps, point at their own model provider, and extend with custom skills.
Who should skip it
Non-technical users and teams who want a managed product with vendor support, or anyone who would rather not maintain their own setup and pay separately for model usage.
Strengths
- Runs locally and stores config and history on your own machine, keeping data private
- Works inside chat apps you already use rather than a separate web app
- Model-agnostic — 40+ bundled providers, subscription sign-in, or local models
- Extensible through a simple file-based skills system (SKILL.md)
- Free and open source with no vendor lock-in
Weaknesses
- Requires technical setup and command-line comfort
- You pay for and manage your own model usage
- No hosted tier or paid support
- Quality and reliability depend on the external model you choose
OpenClaw pricing
Open Source
Free
- Full features, self-hosted on your own machine
- MIT-licensed; stewarded by the OpenClaw Foundation
- No enterprise edition and no paid version
- Bring your own model provider, subscription, or local model
Free tier limits: Free to download and run. You pay your chosen model provider for usage, or use local models.
Note: OpenClaw itself is free and open source; the only ongoing cost is the external LLM you connect it to.
Technical specs
Modalities
Text, Voice
Available models
Where OpenClaw excels
Running a private assistant inside Telegram or Slack
Because OpenClaw connects to messaging apps you already use, you can chat with your assistant where you work without adopting a new interface.
Automating recurring personal tasks on a schedule
Its built-in cron tool lets the assistant run jobs (reminders, lookups, summaries) at set times instead of only responding when prompted.
Extending the assistant with custom skills
The SKILL.md system lets you teach OpenClaw new, repeatable workflows specific to your needs without modifying the core project.
OpenClaw vs. competitors
OpenClaw vs. Manus
Manus is a hosted autonomous agent you give a goal and let run; OpenClaw is self-hosted and works through your own chat apps and model key, trading convenience for control and privacy.
OpenClaw vs. Lindy
Lindy is a managed, no-code assistant and automation platform; OpenClaw is open source and self-hosted, better suited to users who want full ownership rather than a turnkey SaaS.
OpenClaw vs. ChatGPT
ChatGPT is a polished hosted assistant running OpenAI's own models; OpenClaw is a local, model-agnostic assistant you operate from messaging apps with your own API key.
Frequently asked questions
What is OpenClaw?
OpenClaw is a free, open-source personal AI assistant that runs locally and operates through messaging apps like Telegram, WhatsApp, Slack, Discord, Signal, and iMessage. It connects to a model you choose and can browse the web, run scheduled tasks, and use tools on your behalf.
Is OpenClaw free?
Yes. OpenClaw is free and open source under the MIT license, with no enterprise edition or paid version. The only cost is the model you connect it to — an API key, a subscription, or local models on your own hardware.
Which AI models does OpenClaw work with?
OpenClaw is model-agnostic. Its 40+ bundled provider plugins include Anthropic, OpenAI, Google Gemini, DeepSeek, xAI, Mistral, OpenRouter, Groq, and GitHub Copilot, and it also supports subscription sign-in, gateways, custom endpoints, and local inference servers.
How is OpenClaw different from ChatGPT or Manus?
ChatGPT and Manus are hosted products with their own infrastructure. OpenClaw is self-hosted and runs inside your existing messaging apps using your own model key, which gives you more privacy and control but requires technical setup and ongoing maintenance.
Is my data private with OpenClaw?
Configuration and conversation history are stored locally on your own machine, so your data stays under your control. Note that prompts you send to your chosen external model provider are still subject to that provider's data policies.
Integrations & fit
Manus
Professionals and teams who need AI to complete defined tasks end-to-end — research reports, data extraction, scheduled automation, presentations — rather than assist with each step manually
Freemium
Lindy
Individual professionals and small teams who want an AI assistant that works across their actual email, calendar, and CRM without building explicit automation flows
Paid
AgentScope
Technical users and small teams who want a self-hosted assistant in DingTalk, Feishu, WeCom, WeChat, or QQ, offline use with small local models, and built-in tool guards with an OS-level shell sandbox
Free
TinyHumans
Individuals and solo operators on macOS, Windows, or Linux who want a desktop agent that keeps a local, readable memory of their email, chat, and work apps and can act across them, without wiring API keys or running a server
FreemiumAbout OpenClaw
OpenClaw, created by Peter Steinberger and now stewarded by the OpenClaw Foundation (an independent US 501(c)(3) non-profit that employs the core team and signs every release), is a personal AI assistant designed to live inside your existing chat apps rather than a separate web app. It is model-agnostic: 40+ bundled provider plugins cover Anthropic, OpenAI, Google, DeepSeek, xAI, Mistral, OpenRouter, and others, and you can also use subscription sign-in, gateways, or local models. Instead of a single chatbot, it routes inbound messages from different channels and accounts to isolated agents with their own workspaces and sessions, and extends itself through a file-based 'skills' system (directories containing a SKILL.md with instructions and tool usage). Configuration and conversation history are stored locally, which keeps data on your own machine and lets behavior persist across sessions. First-class tools include a browser, canvas, cron scheduling, and chat-platform actions, plus optional voice wake and talk modes. With roughly 390,000 GitHub stars (September 2026), it is among the most-starred software repositories on GitHub. It installs with a one-line installer or npm, pnpm, Docker, or Nix, and has desktop apps for macOS, Windows, and Linux plus iOS and Android nodes. OpenClaw makes the most sense for developers and tinkerers who want full control and privacy over a hosted assistant; it is not a turnkey product, and you are responsible for setup, model costs, and oversight of any actions it takes.
Are you the founder? Claim this listing →