OpenClaw is an open-source personal AI assistant that runs locally and connects to over 13 messaging platforms including WhatsApp, Telegram, Slack, Discord, Signal, iMessage, and Microsoft Teams. With over 202,000 GitHub stars and MIT licensing, it has become one of the most popular AI projects on GitHub, offering a unified control plane for managing AI interactions across every major communication channel. ## Why a Multi-Platform AI Assistant Matters Most AI assistants are locked inside a single interface. ChatGPT lives in its web app, Claude lives in its web app, and Copilot lives in its editor. Users who want AI assistance across their actual communication channels, where conversations happen with colleagues, friends, and clients, are forced to copy-paste between windows or settle for fragmented experiences. OpenClaw solves this by creating a single AI gateway that plugs into every messaging platform a user already relies on. The assistant follows the user across channels rather than requiring the user to come to the assistant. ## Architecture and Design ### Local Gateway At the core of OpenClaw is a local Gateway server that acts as the central control plane. This Gateway manages sessions, channels, tools, and events through a WebSocket-based architecture. All processing happens on the user's machine, with AI model calls routed through the user's own API keys. There is no intermediary server collecting conversation data. The Gateway supports both daemon mode for always-on operation and foreground mode for development and debugging. Installation is as simple as running npm install -g openclaw@latest followed by the onboarding command. ### Multi-Channel Inbox OpenClaw's defining feature is its unified inbox spanning 13 messaging platforms. WhatsApp, Telegram, Slack, Discord, Google Chat, Signal, iMessage via BlueBubbles, Microsoft Teams, Matrix, Zalo, and WebChat are all supported. Each channel connects to the same underlying AI agent, maintaining context and memory across platforms. This means a conversation started on Telegram can be continued on Slack without losing context. The AI remembers what was discussed regardless of which channel delivered the message. ### Voice Capabilities The platform includes built-in voice support with two modes. Always-on speech mode provides continuous listening and response, while push-to-talk mode offers more controlled interactions. Voice input is transcribed and processed through the same AI pipeline as text messages. ### Live Canvas The A2UI Live Canvas provides a visual workspace where the AI can render interactive content, charts, code previews, and collaborative documents in real time. This goes beyond simple text responses, enabling rich visual interactions when the use case demands it. ### Skills Platform OpenClaw supports extensibility through a three-tier skills system. Bundled skills ship with the platform and cover common tasks. Managed skills are installed from a registry. Workspace-level skills are custom extensions built by the user for specific workflows. ### Multi-Agent Routing For advanced users, OpenClaw supports routing messages to isolated workspaces and sessions with different AI configurations. This enables running specialized agents for different domains, such as a coding agent, a writing agent, and a research agent, all accessible through the same messaging channels. ## Automation Features OpenClaw includes built-in automation through cron jobs, webhooks, and Gmail Pub/Sub integration. Users can schedule recurring AI tasks, trigger actions based on external events, and automate email processing without additional tools. Browser control is another automation capability. OpenClaw can drive a dedicated Chrome or Chromium instance, enabling web scraping, form filling, and automated browsing tasks directed by the AI. ## Model Recommendation The official documentation recommends Anthropic Pro/Max with Opus 4.6 as the primary model, though OpenClaw works with any LLM provider that exposes a compatible API. Users can configure different models for different workspaces or switch models on the fly. ## Device Node Support OpenClaw extends beyond desktop with device nodes for macOS, iOS, and Android. These nodes allow the AI to interact with device-specific features and sensors, bringing the assistant into mobile contexts. ## Security DM pairing policies ensure that the AI only responds to authorized users. The security model defaults to restrictive permissions, requiring explicit configuration to enable sensitive capabilities like browser control or file system access. ## Installation OpenClaw is available through npm with stable, beta, and dev release channels. The stable channel receives tested releases, beta provides prerelease builds, and dev tracks the moving head of the main branch. Node.js 22 or later is required. ## Limitations The breadth of platform integrations means that some channels are more polished than others. iMessage support requires a Mac running BlueBubbles, which adds complexity. Voice features depend on local audio hardware and may not work in all environments. The project's rapid growth has outpaced documentation in some areas, and configuration for advanced features like multi-agent routing requires familiarity with the architecture. ## Market Position With over 202,000 stars, OpenClaw is one of the most starred AI projects on GitHub. Its multi-platform approach differentiates it from single-interface assistants, and the local-first architecture appeals to users who want AI assistance without surrendering their conversation data to third-party servers.