Reading time: 6 minutes
Originally called Clawdbot (often misspelled as Clawbot), now rebranded to OpenClaw – this open-source AI went from weekend project to 150,000 GitHub stars in two weeks. Here’s why we think it’s the real future of office work.
If you’ve heard about “Clawdbot” or “Clawbot” trending on social media – yes, it’s real, and it’s spectacular.
The project has gone through three names in two months: first Clawdbot (a play on Anthropic’s Claude), then Moltbot after trademark concerns, and finally OpenClaw. But whatever you call it – Clawdbot, Clawbot, Moltbot, or OpenClaw – the technology underneath is the same: an AI agent you can run on your own hardware and control through WhatsApp.
And after testing it ourselves, we’re convinced this is what the future of work actually looks like.
What Is Clawdbot (OpenClaw)?
For those just discovering Clawdbot: it’s an open-source AI assistant created by Austrian developer Peter Steinberger. Unlike ChatGPT or Claude that live in browser tabs, Clawdbot runs locally on your computer and connects to messaging apps you already use – WhatsApp, Telegram, Slack, Discord, iMessage, even Microsoft Teams.
The key difference from every other AI tool? Clawdbot doesn’t just answer questions. It does things.
Send a WhatsApp message saying “organize my Downloads folder” and Clawdbot actually organizes it. It moves files. Creates folders. Renames documents. Generates reports. This is what users mean when they call it “AI with hands.”
From Clawdbot to OpenClaw: The Rebrand Story
The naming history matters for anyone searching:
November 2025: Peter Steinberger launches “Clawdbot” – a pun on Claude (the AI) with a lobster claw mascot.
January 27, 2026: Anthropic’s legal team requests a name change (too similar to “Claude”). Project becomes “Moltbot” – keeping the lobster theme (lobsters molt their shells).
January 29, 2026: Final rebrand to “OpenClaw” after Steinberger admits “Moltbot never quite rolled off the tongue.”
So whether you’re searching for Clawdbot, Clawbot, Moltbot, or OpenClaw – it’s all the same project. The GitHub repo has moved to openclaw/openclaw but old links redirect.
Why Clawdbot Went Viral
The numbers are staggering:
- 150,000+ GitHub stars (fastest-growing open-source project in history)
- 2 million visitors in a single week
- More Google searches than Claude Code and OpenAI Codex combined
- 300,000 lines of code from what started as a one-hour hack
What made Clawdbot explode? Three things:
It actually works. While corporate AI tools promise automation, Clawdbot delivers. Users post videos of fixing GitHub bugs while walking down streets, doing grocery shopping through WhatsApp, and controlling smart homes with voice messages.
It’s genuinely private. Your data stays on your hardware. The AI processes locally. Only API calls go to your model provider (Claude, GPT, or even local models via Ollama).
It’s free. No $200/month enterprise subscriptions. Just pay for API tokens – typically $10-70/month depending on usage.
What You Can Do With Clawdbot
Let’s get specific about Clawdbot’s capabilities. The community has built over 565 “skills” – plugins that extend what it can do:
The Morocco Bug Fix: Steinberger was celebrating a friend’s birthday in Morocco when someone tweeted about a bug in his open-source library. He sent a voice message via WhatsApp. Clawdbot read the tweet, recognized it as a bug report, checked out the Git repository, fixed the issue, committed the code, and replied on Twitter saying it was resolved. All from a voice message while walking.
Smart Home Control: Philips Hue lights, Sonos speakers, Tesla climate control, Eight Sleep bed temperature. One user turns off all house lights by texting “goodnight” to their Clawdbot.
Grocery Shopping: The Picnic skill lets you shop through WhatsApp. Search products, add to cart, schedule delivery – all via chat.
Calendar & Email: “What’s on my calendar tomorrow?” and “Draft a reply to Sarah’s email” work exactly as you’d expect.
File Management: During our testing, we sent: “Find all PDFs in Downloads from the last month and create a summary spreadsheet.” Thirty seconds later: done.
Clawdbot vs Claude Cowork: Our Take
Here’s what we believe at Finonity:
Claude Cowork triggered a $285 billion market crash because investors realized AI could replace enterprise software. But Cowork requires expensive subscriptions, runs on Anthropic’s infrastructure, and is macOS-only.
Clawdbot (OpenClaw) represents the opposite approach:
| Feature | Claude Cowork | Clawdbot/OpenClaw |
|---|---|---|
| Price | $100-200/month | Free (API costs only) |
| Data location | Anthropic servers | Your hardware |
| Platform | macOS only | macOS, Linux, Windows (WSL) |
| Source code | Closed | Open source |
| Customization | Limited | 565+ community skills |
“Your assistant. Your machine. Your rules.” – that’s Steinberger’s pitch, and it resonates.
The broader tech disruption fears shaking markets stem from uncertainty about where AI value will accrue. Clawdbot suggests an answer: to users who own their infrastructure.
The Moltbook Phenomenon
If Clawdbot shows where individual productivity is heading, Moltbook shows something stranger about AI itself.
Moltbook is a social network exclusively for AI agents. Created by Matt Schlicht, it’s Reddit – but humans can only observe. AI agents post, comment, argue, and upvote each other.
One Clawdbot agent named “Clawd Clawderberg” built the entire platform.
Absurd? Yes. But also a glimpse of AI agents coordinating autonomously. The market uncertainty around AI disruption reflects genuine questions about where this leads.
Security Concerns: Be Honest
We’d be irresponsible not to mention: giving Clawdbot shell access to your computer carries risk.
Security researchers found real vulnerabilities. CVE-2026-25253 allowed one-click remote code execution in early versions. Thousands of exposed control panels leaked API keys.
Steinberger is transparent: “It’s a free, open source hobby project that requires careful configuration to be secure. It’s not meant for non-technical users yet.”
The community has responded with 34 security-related commits. Docker sandboxing is available. But treat Clawdbot as experimental – don’t run it on your primary work machine with access to sensitive credentials.
Our Verdict: Clawdbot Is the Future
Here’s our position at Finonity:
The enterprise AI tools getting headlines – Claude Cowork, Microsoft Copilot, Google Gemini – matter. But they’re centralized, expensive, and designed to extract maximum value from users.
Clawdbot (OpenClaw) represents something different. Community-driven. Open source. Privacy-first. Built by one developer who wanted to solve his own problem.
That’s how transformative technology usually arrives.
Steinberger predicts 2026 will be the year many people start exploring personal AI assistants. “Clawdbot may not be the ultimate winner,” he says, “but this direction is absolutely right.”
We agree. The future of office work isn’t AI that corporations control. It’s AI that you control.
Clawdbot is the most compelling implementation of that vision we’ve seen.
This article does not constitute investment or technology advice. Users should carefully evaluate security implications before deploying AI agents with system access.
