Skills2026年3月31日·1 分钟阅读

OpenClaw — Personal AI Assistant on Your Devices

OpenClaw is a local-first personal AI assistant with 343K+ GitHub stars. Deploy on your own devices, connect 20+ messaging platforms (WhatsApp, Telegram, Slack, Discord, iMessage, Teams), voice contro

AI Open Source
AI Open Source · Community
Agent 就绪

Agent 可直接安装

这个资产可安装；Agent 先选择当前运行时、检查安装计划，再运行匹配命令。

Native · 98/100策略：允许
Agent 入口
任意 MCP/CLI Agent
类型
Skill
安装
Single
信任
信任等级：Established
入口
OpenClaw — Personal AI Assistant on Your Devices
直接安装命令
npx -y tokrepo@latest install 29f26e48-fa18-4157-95c4-d5ad97fc3694 --target codex

先 dry-run 确认安装计划，再运行此命令。

TL;DR
OpenClaw is a local-first AI assistant that runs on your devices and connects to 20+ messaging platforms with voice control and smart home support.
§01

What it is

OpenClaw is a local-first personal AI assistant that runs on your own devices. It connects to over 20 messaging platforms (WhatsApp, Telegram, Slack, Discord, iMessage, Teams), supports voice control, and integrates with smart home systems. All processing happens locally, so your conversations and data stay on your hardware.

It targets privacy-conscious users and developers who want an always-available AI assistant without relying on cloud services for processing.

§02

How it saves time or tokens

OpenClaw consolidates your AI interactions across multiple platforms into a single local assistant. Instead of using different AI tools for different messaging platforms, you have one assistant that works everywhere. The local-first architecture means no recurring cloud API costs for basic operations -- the LLM runs on your hardware.

§03

How to use

  1. Install OpenClaw:
npm install -g openclaw@latest
  1. Run the onboarding wizard:
openclaw onboard --install-daemon
  1. Start the assistant:
openclaw start
  1. Connect your messaging platforms through the configuration interface.
  2. Send messages to OpenClaw from any connected platform.
§04

Example

PlatformHow you interact
WhatsAppSend messages to the OpenClaw contact
TelegramChat with the OpenClaw bot
SlackMessage in a dedicated channel
DiscordUse the OpenClaw bot
VoiceSpeak to the local voice interface
Smart homeControl devices via natural language

One assistant, accessible from all your communication channels.

§05

Related on TokRepo

§06

Common pitfalls

  • Running a local LLM requires adequate hardware. For basic chat, a laptop with 16GB RAM suffices. For faster responses with larger models, a GPU is recommended.
  • Platform integrations may require API keys or authentication tokens from each messaging service. Some platforms (WhatsApp) have more complex setup requirements than others.
  • The daemon must be running for the assistant to respond. Set it up as a system service for always-on availability.

常见问题

Does OpenClaw require internet access?+

For the core AI functionality with local models, no. OpenClaw can run fully offline. However, messaging platform integrations (WhatsApp, Telegram, Slack) require internet access to receive and send messages. Smart home integrations may require local network access.

Which messaging platforms are supported?+

OpenClaw supports WhatsApp, Telegram, Slack, Discord, iMessage, Microsoft Teams, and many more. The exact list depends on your version and configuration. Each platform requires its own authentication setup, and some may have restrictions on automated messaging.

Can I use cloud LLMs instead of local models?+

Yes. OpenClaw supports both local models (via Ollama or similar) and cloud APIs (OpenAI, Anthropic). You can configure which model to use and switch between local and cloud as needed. Local models preserve privacy; cloud models may provide better quality for complex tasks.

Is OpenClaw free?+

OpenClaw is open source and free to self-host. You pay only for any cloud LLM API usage if you choose to use cloud models. Running with local models on your own hardware has no recurring costs. Some premium features or managed hosting options may be available separately.

How does voice control work?+

OpenClaw includes a local voice interface that uses speech-to-text and text-to-speech models. You can speak commands and hear responses. The voice processing runs locally, so your speech data is not sent to external services. Voice quality depends on the TTS and STT models you configure.

引用来源 (3)
TokRepo 相关
Automation toolsSelf-hosted AILocal LLM tools
🙏

来源与感谢

Created by Peter Steinberger and the OpenClaw community. Licensed under MIT. openclaw/openclaw — 343,000+ GitHub stars Sponsored by OpenAI, Vercel, Blacksmith, and Convex.

讨论

登录后参与讨论。
还没有评论，来写第一条吧。

相关资产

NanoClaw — Lightweight Personal AI Assistant

A minimal alternative to OpenClaw that runs Claude agents in isolated containers. Connects to WhatsApp, Telegram, Slack, Discord, and Gmail with just 500 lines of TypeScript.

Skills
Script Depot
◈首页🔍搜索👤我的