What this is
OpenClaw is the local-first personal AI assistant project that went from a public repository in November 2025 to 372,000 GitHub stars and 77,000 forks in roughly six months. It is sponsored by OpenAI, GitHub, NVIDIA, and Vercel, runs on any OS, and integrates with 20+ messaging channels. This page is a 2026-05-15 snapshot of the project, its scope, and what its growth says about where personal AI assistants are going.
OpenClaw at a Glance
| Metric | Value (2026-05-15) |
|---|---|
| GitHub stars | 372,033 |
| Forks | 77,021 |
| Commits on main | 49,692 |
| Repository created | 2025-11-24 |
| Latest release | v2026.5.12 (2026-05-14) |
| Primary language | TypeScript |
| Runtime | Node.js 24 (or 22.16+) |
| Supported OS | macOS, Linux, Windows (WSL2), iOS, Android |
| License | MIT |
| Stated sponsors | OpenAI, GitHub, NVIDIA, Vercel |
Channel Integrations (selected)
| Channel | Direction |
|---|---|
| In + Out | |
| Telegram | In + Out |
| Slack | In + Out |
| Discord | In + Out |
| Signal | In + Out |
| iMessage (macOS bridge) | In + Out |
| SMS (Twilio bridge) | In + Out |
| Email (IMAP/SMTP) | In + Out |
| Voice (macOS, iOS, Android) | In + Out |
| Browser canvas | Live |
| Cron / automation | Out |
Architectural Highlights
- Local-first gateway. A single OpenClaw gateway runs on the user's hardware; channels connect to it rather than to a SaaS endpoint.
- Multi-agent routing. Requests are dispatched to specialised agents (browser, canvas, calendar, code, automation).
- Skill registry. Skills are TypeScript packages registered at runtime, distributed via npm or GitHub.
- Bring-your-own-model. Defaults to local Ollama; can fall back to OpenAI, Anthropic, Gemini, or other providers via per-skill routing.
- Companion apps. macOS, iOS, and Android companion apps for voice and notifications; the main runtime stays on the user's machine.
Six Things the OpenClaw Growth Tells You
- 372K stars in six months is unprecedented for an MIT-licensed AI assistant. It outpaces every prior local-LLM tool (Ollama, LM Studio, Open Interpreter) at similar age.
- The sponsor list signals platform validation. OpenAI sponsoring a project that defaults to local inference is the most interesting strategic signal.
- 20+ channel coverage moves personal AI past chat-window UX. The "your AI lives in your existing inboxes" pattern is the architecture that scales beyond power users.
- Skill-registry pattern mirrors Claude Skills and MCP. The convergence on packaged-skill ecosystems is the 2026 platform layer.
- TypeScript + Node.js stack lowers the contributor bar. 49K commits + 77K forks reflect the easiest plug-in surface in the local-AI category.
- "The lobster way" branding works. Identity-rich open-source projects compound network effects faster than generic ones; OpenClaw is the 2026 case study.
What This Means for AI Visibility
OpenClaw is the first credible candidate for a default personal AI assistant that does not route everything through ChatGPT, Claude, or Gemini. Brands that want to be present inside agentic workflows running in OpenClaw need their content reachable to the local-skill layer (well-structured pages, MCP server, llms.txt, schema.org) just as much as they need it reachable to ChatGPT. Skill authors are the new discovery channel; brands without a presence in their tooling reduce surface area.
Methodology
Stats sourced live from the OpenClaw GitHub repository on 2026-05-15 via the GitHub API. Channel and architecture details are drawn from the project's homepage and README; sponsor list is the publicly stated set in the repository metadata.
How Presenc AI Helps
Presenc AI tracks brand mentions and product recommendations across the surfaces OpenClaw users actually consult: WhatsApp answer threads using local models, Slack/Discord channel responses, voice queries on iOS, and the canvas browser interactions. Skill authors and brands that need to know whether OpenClaw users see them inside its multi-agent routing use Presenc to monitor and benchmark this surface alongside ChatGPT, Claude, Gemini, and Perplexity.