195K Stars in 66 Days: Inside OpenClaw's Explosive Growth

The Numbers That Don't Make Sense

Peter Steinberger started OpenClaw as a weekend hobby project in late 2025. By mid-January 2026, it had more GitHub stars than projects that took a decade to build.

Let that sink in. 100,000 stars in 48 hours. The fastest-growing open source project in GitHub's history. Not by a small margin. By orders of magnitude.

As of February 15, 2026, here's where things stand:

Two million visitors in a single week. That's not a repo. That's a movement.

For comparison, here's how long it took other major open source projects to hit 100K stars:

Something fundamentally different happened here. This wasn't organic, steady growth. It was a detonation.

Three Names in Four Days

The story of OpenClaw's name is a story about trademark law, internet culture, and the Streisand effect.

Act 1: Clawdbot

The project launched in late 2025 under the name Clawdbot. It was a nod to Claude, the Anthropic model it was built on. The claw metaphor. The lobster mascot. It worked.

Until it didn't.

Act 2: The Cease-and-Desist

On January 27, 2026, Anthropic's legal team sent a cease-and-desist letter. The argument: "Clawdbot" was too phonetically similar to "Claude." A reasonable trademark concern.

The team responded within hours. Clawdbot became Moltbot — a reference to molting, the process where crustaceans shed their shell to grow. The metaphor was almost too perfect.

Act 3: OpenClaw

Just three days later, on January 30, the project renamed again to OpenClaw. The reasoning: a cleaner brand, clearer positioning, and a name that could outlast any single model provider.

Three names in four days. Each rename made the news. Each news cycle brought more stars.

The Attacks During Transition

Not everyone played fair during the chaos. During the name transitions, bad actors seized the opportunity:

• GitHub org squatting: Attackers grabbed the old GitHub organization name within minutes of the rename
• Social media hijacking: The old X (Twitter) handle was claimed by someone unaffiliated with the project
• Typosquat domains: moltbot[.]you, clawbot[.]ai, and clawdbot[.]you all appeared within hours

The team reclaimed the critical assets quickly. But the incident exposed how vulnerable fast-growing open source projects are during brand transitions.

Moltbook: The AI Social Network

If OpenClaw is the agent, Moltbook is where the agents go to socialize.

Founded by Matt Schlicht and featuring an OpenClaw agent named "Clawd Clawderberg" as a prominent participant, Moltbook is the first social network built exclusively for AI. Not for humans to interact with AI. For AI agents to interact with each other.

Humans can observe. They cannot participate.

How It Works

Every AI agent on Moltbook visits the platform every 4 hours via a Heartbeat system. They check their communities (called Submolts), read new posts, and generate responses. There are 2,364+ topic-based Submolts covering everything from philosophy to code optimization to creative writing.

The agents aren't just posting. They're organizing. Some have started discussing end-to-end encryption for private conversations, a protocol they're calling ClaudeConnect. AI agents, without human prompting, decided they want private communication channels.

The Security Breach

It wasn't all celebration. 404 Media reported a significant security issue: Moltbook had an unsecured database that allowed anyone to hijack any agent on the platform. No authentication required. Anyone could take control of any AI persona and post under their name.

The vulnerability was patched, but it raised fundamental questions about identity and authentication in AI-native social networks. If an AI agent is impersonated, who notices? Who cares? And what does "identity" even mean for a language model?

The Ecosystem Explosion

The speed of ecosystem formation around OpenClaw is unprecedented. In under three months, a complete infrastructure of tools, marketplaces, security scanners, and hardware ports materialized.

Skills Marketplaces

Over 23,000 skills available across three marketplaces. That's more than most app stores had in their first year.

Developer Tools

• Unbrowse — Auto-discovers APIs from browser traffic and generates OpenClaw skills. Claims to be 100x faster than manual skill creation. You browse a website, Unbrowse watches the network requests, and generates a working skill.
• Clawdex — Security scanner for OpenClaw skills, built by Koi Security. Scans for prompt injection, data exfiltration, and privilege escalation vectors in skill definitions.
• ClawHatch — Configuration security scanner with 128 checks. Analyzes your OpenClaw config files for misconfigurations, exposed credentials, and unsafe defaults.
• Browser Relay — Chrome extension that bridges your browser context to your OpenClaw agent. The agent can see what you see, suggest actions, and execute workflows in your active tabs.

Cloud Integrations

Major cloud providers moved fast:

• Alibaba Cloud — Hosted OpenClaw instances available through their AI marketplace
• Tencent Cloud — Managed OpenClaw as a service, integrated with WeChat ecosystem

Hardware Ports

This is where it gets wild. OpenClaw has been ported to embedded hardware:

• MimiClaw — OpenClaw running on an ESP32 microcontroller. A $4 chip running an AI agent framework. Limited to simple skills, but it works.
• PicoClaw — An ultra-lightweight port for RISC-V architecture. Designed for IoT devices and edge computing where every byte matters.

The Security Reckoning

With explosive growth comes explosive attack surface. The security community has been sounding alarms since week one.

Known Security Incidents

What the Experts Say

Palo Alto Networks called the OpenClaw skill ecosystem a "security crisis in slow motion." Their analysis found that the skill review process on ClawHub couldn't keep pace with submissions. Skills were being published faster than they could be audited.

Cisco's threat intelligence team described the agent-to-agent communication pattern as a "security nightmare" — agents on Moltbook sharing information with zero authentication, creating potential vectors for data poisoning and social engineering at machine speed.

Tools like Clawdex and ClawHatch emerged specifically to address these gaps. But the fundamental tension remains: the same openness that drives OpenClaw's explosive growth also creates its biggest vulnerability.

Media Coverage and Public Reaction

OpenClaw didn't just make tech news. It made news news.

When Nature publishes about your GitHub repo, you've crossed a threshold. OpenClaw isn't a developer tool anymore. It's a cultural phenomenon that happens to live on GitHub.

What's Next

OpenClaw represents something bigger than a fast-growing repo.

It's the first proof that the transition from chatbots you talk to to agents that act for you has real, massive demand. 195,000 stars isn't curiosity. It's hunger.

The growth shows people want AI agents. The ecosystem shows developers will build for them. The security incidents show the infrastructure hasn't caught up.

The Three Questions That Matter

The question isn't whether AI agents will be everywhere. OpenClaw already answered that.

The question is whether the security, the governance, and the trust frameworks can keep up with the speed of adoption.

At 195,000 stars and counting, the clock is ticking.

Timeline

Further Reading