OpenClaw
Type: full-code · Vendor: OpenClaw · Language: TypeScript · License: MIT · Status: active · Status in practice: emerging · First released: 2025-11-24
Run a personal AI assistant on your own devices that listens and replies across the chat and voice channels you already use (WhatsApp, Telegram, Slack, Discord, iMessage, WeChat, Matrix, …), via a self-hosted Gateway, Workspace, Channels, and Skills.
Description. OpenClaw is the MIT-licensed TypeScript personal AI assistant from the openclaw project. Per its README, 'OpenClaw is a personal AI assistant you run on your own devices. It answers you on the channels you already use. It can speak and listen on macOS/iOS/Android, and can render a live Canvas you control. The Gateway is just the control plane — the product is the assistant.' Distinguishing features: 22+ chat/voice channels listed (WhatsApp, Telegram, Slack, Discord, Google Chat, Signal, iMessage, IRC, Microsoft Teams, Matrix, Feishu, LINE, Mattermost, Nextcloud Talk, Nostr, Synology Chat, Tlon, Twitch, Zalo, WeChat, QQ, WebChat), CLI onboarding via `openclaw onboard`, and OpenAI/GitHub/NVIDIA/Vercel as listed sponsors. Distinct project from Gen-Verse/OpenClaw-RL (which is the complementary RL training framework that wraps a model in the OpenClaw plugin).
Agent loop shape. Gateway control plane fronting a Workspace; per-channel adapters (chat apps + voice) feed user messages into an assistant runtime that calls model providers and Skills, optionally renders a live Canvas, and responds back on the source channel. Self-hosted via npm/pnpm/bun on macOS, Linux, or Windows (WSL2).
Primary use cases
- self-hosted multi-channel personal AI assistant
- voice + text chat across messaging platforms (WhatsApp/Telegram/Slack/Signal/iMessage/WeChat)
- Skills-based extensibility on top of a local Gateway
Key concepts
- Gateway (docs) — The control plane in front of the assistant runtime; routes incoming channel traffic into the workspace.
- Workspace (docs) — Container for the assistant's state, skills, and channel bindings; configured via `openclaw onboard`.
- Channels (docs) — Per-platform adapters (WhatsApp, Telegram, Slack, Signal, iMessage, WeChat, Matrix, etc.) that connect the assistant to existing messaging tools.
- Skills → agent-skills (docs) — Pluggable capabilities the assistant can invoke; configured during onboarding.
- Canvas (docs) — A live, user-controllable visual surface the assistant can render alongside text/voice responses.
Patterns this full-code implements —
- ★★Event-Driven Agent
Incoming chat / voice messages from connected channels are the event source driving the assistant loop; no scheduled or batch invocation.
- ★Multilingual Voice Agent Stack
Voice in and out on macOS/iOS/Android is a documented headline feature.
- ★Agent Skills
Skills are a core onboarding-configured capability of the workspace.
- ★★Session Isolation
Self-hosted single-user by design; per-channel context maintained by the Gateway. No formal multi-tenant session-isolation API documented.
Neighbourhood
Click any neighbour to follow the lineage. Scroll to zoom, drag to pan.
Alternatives & relatives
- no-code · visual builder
Lindy
Hosted personal-AI-employee SaaS; OpenClaw is the self-hosted, multi-channel counterpart.
- app · closed product
Dust
Hosted AI assistant; different deployment model.
- full-code · framework
Same brand and openclaw.ai homepage; OpenClaw-RL is the RL training framework that wraps a model in the OpenClaw plugin so OpenClaw can intercept and learn from conversations.
Listed as alternative by (3)
References
Provenance
- Last analyzed:
- Last updated:
- Verification status: partial