Clawdbot → Moltbot → OpenClaw: The Full History of the Viral AI Agent

The project that became OpenClaw started as a side project called Clawdbot. The original concept was simple: a personal AI assistant that you could control through WhatsApp, running on your own machine. The first version was rough — a Python script, a Telegram bot, and a lot of duct tape.

Clawdbot gained traction slowly at first, then all at once. A post on Hacker News in late 2025 sent it viral. The GitHub star count went from a few hundred to tens of thousands in a week. The creator, overwhelmed by the response, briefly rebranded to Moltbot while rethinking the architecture.

The Moltbot phase was short. The project was rewritten in TypeScript, the architecture was overhauled, and the name changed again — this time to OpenClaw. The new name reflected the project's ambition: an open, extensible platform for personal AI agents, not just a WhatsApp bot.

OpenClaw's growth from there was extraordinary. The Skills system — a plugin architecture that let users extend the agent with new capabilities — created a community of contributors. The ClawHub marketplace launched with dozens of community-built integrations. The GitHub star count crossed 100,000, then 200,000.

The viral growth also brought scrutiny. The CVE-2026-25253 security disclosure in February 2026 was a wake-up call for the community. The project's single-user architecture, designed for personal use, wasn't built for the scale of deployment it had achieved.

Today, OpenClaw is the foundation of an ecosystem. Managed hosting providers like MyClaw and Talking Claw have built products on top of it. Lightweight alternatives like NanoClaw, ZeroClaw, and PicoClaw have emerged for users who want something smaller and more auditable. The project that started as a WhatsApp bot has become the defining platform of the personal AI agent movement.