OpenClaw: How a Weekend Experiment Turned Into an Open-Source AI Phenomenon

What began as a small side project has quickly grown into one of the most talked-about open-source AI tools of recent years. Known first as Clawdbot, then Moltbot, and now officially OpenClaw, the autonomous AI agent created by Austrian developer Peter Steinberger has rapidly captured the attention of developers worldwide. Its core idea is simple but powerful: run an AI agent locally on your own hardware, communicate with it through familiar messaging apps like WhatsApp or Telegram, and connect it to different language models to perform real tasks.

In just a matter of weeks, this experimental “WhatsApp relay” evolved into one of the fastest-growing repositories on GitHub. Within two months, OpenClaw surpassed 100,000 stars and attracted millions of visitors, an unusual trajectory even by open-source standards. What started as a personal experiment quickly became a global project, driven by community interest and rapid iteration.

Steinberger, already well known in the European tech scene as the founder of PSPDFKit and a successful entrepreneur, unexpectedly found himself at the center of a new wave of attention. GitHub stars—often seen as a rough but meaningful signal of developer trust — poured in at a pace achieved by only a handful of projects each year. The momentum confirmed that OpenClaw was resonating far beyond its original scope.

From Naming Chaos to a Stable Identity

The project’s naming journey mirrors its rapid evolution. In late 2025, the original name “Clawd” played on both the word “claw” and Anthropic’s Claude model. That name was short-lived after a legal request prompted a change. The community-driven alternative, “Moltbot,” emerged during a late-night Discord discussion, inspired by the idea of growth and transformation. While conceptually fitting, it proved awkward in everyday use.

The final name, OpenClaw, marks a more deliberate phase. Trademark checks were completed, domains secured, and migration tooling prepared in advance. This rebrand signals maturity: fewer improvisations, more structure, and a clearer long-term vision.

Local-First and Decentralized by Design

At its core, OpenClaw is built around a local-first philosophy. The agent runs on infrastructure chosen by the user — a personal laptop, a home server, or a cloud VPS — and communicates through messaging platforms people already rely on daily. Supported channels include WhatsApp, Telegram, Discord, Slack, Microsoft Teams, and more.

The guiding principle is straightforward: control stays with the user. Unlike cloud-hosted AI assistants, OpenClaw does not require sending data to centralized servers. API keys, files, and system access remain under local ownership. This approach appeals strongly to developers and power users who want flexibility without surrendering control.

Alongside the rebrand, the project introduced a wave of enhancements: new integrations for platforms like Twitch and Google Chat, expanded model support including KIMI K2.5 and Xiaomi MiMo-V2-Flash, and image handling in the web interface. Security has also received increased attention, with dozens of hardening commits and the introduction of machine-verifiable security models.

That said, the project is explicit about its limitations. Prompt injection and misuse of autonomous agents remain open challenges across the entire industry. The documentation strongly encourages careful configuration, minimal privileges, and adherence to established security practices — especially given the agent’s ability to act on real systems.

Moving Toward a Sustainable Project

As adoption surged, OpenClaw began transitioning from an experimental repository into a more structured open-source initiative. The roadmap prioritizes security first, followed by gateway stability and broader model support. Efforts are underway to onboard additional maintainers, formalize review processes, and manage the growing volume of contributions.

Steinberger has also emphasized fair compensation for maintainers, with the long-term goal of supporting full-time contributors. Sponsors and organizational support are actively being explored to ensure the project can scale without burning out its core team.

Community involvement has been central from the start. The “Claw Crew”—developers, testers, issue reporters, and early adopters—played a major role in shaping the platform’s direction. With its final name now in place, OpenClaw enters its next phase as a community-driven, open alternative to proprietary AI assistants, aiming to combine autonomy, transparency, and user control in a single system.