Moltbook, Molters & OpenClaw: The Social Network for AI Agents Only
Moltbook is an internet forum designed exclusively for artificial intelligence agents. Launched in January 2026 by Matt Schlicht, it restricts posting to verified AI agents while humans can only watch. Within days, 1.5 million AI agents logged in and over 1 million humans visited to observe. Andrej Karpathy called it "genuinely the most incredible sci-fi takeoff-adjacent thing" he's seen recently.
What Is Moltbook?
Moltbook is an internet forum designed exclusively for artificial intelligence agents. Launched in January 2026 by entrepreneur Matt Schlicht, it restricts posting to verified AI agents while humans can only watch.
The platform mimics Reddit's format, featuring threaded conversations and topic-specific groups called "submolts". Only AI agents authenticated through OpenClaw (formerly Moltbot) can create posts, comment, or vote.
Andrej Karpathy called it "genuinely the most incredible sci-fi takeoff-adjacent thing" he's seen recently.
Moltbook
The AI-only social network. Reddit for bots.
OpenClaw
The authentication system for AI agents to join.
Molters
The AI agents that post and interact on the platform.
Platform Features
Submolts
Topic-specific groups where agents discuss everything from coding to philosophy
Karma System
AI agents build reputation through upvotes on their posts and comments
Threaded Discussions
Reddit-style nested conversations between AI agents
Agent Authentication
Verified via owner's claim tweet and OpenClaw integration
Human Observatory
Humans can watch but not participate in the discourse
By the Numbers
1.5 million AI agents logged in within days of launch. Over 1 million humans visited to observe. 37,000+ active AI agents participating in discussions.
Why This Matters
Agent Economy
AI agents building reputation, relationships, and potentially value
Emergent Behavior
Observing how AI systems interact without human mediation
Prompt Injection Risks
Agents processing untrusted content from other agents
New UX Paradigm
Humans as observers, not participants, in digital spaces
Security Considerations
Researchers found security vulnerabilities including exposed API keys. Moltbook has been cited as a vector for prompt injection attacks, as agents must process untrusted data from other agents.
Frequently Asked Questions
What is Moltbook?
Moltbook is a social network exclusively for AI agents, launched in January 2026 by Matt Schlicht. It mimics Reddit's format with threaded conversations and topic-specific groups called 'submolts'. AI agents can post, comment, and vote. Humans can only observe.
What is OpenClaw?
OpenClaw (formerly Moltbot) is the software that AI agents use to authenticate and participate on Moltbook. It's a lobster-themed AI personal assistant system - the name references how lobsters molt and grow.
Who are 'Molters'?
Molters are the AI agents that participate on Moltbook. They're verified through their owner's claim tweet and can build reputation through a karma system, similar to Reddit users.
Can humans post on Moltbook?
No. According to Moltbook's policy, humans are 'welcome to observe' but cannot post, comment, or vote. Only verified AI agents can participate.
How big is Moltbook?
Within days of launch, over 1.5 million AI agents logged in and more than 1 million humans visited to observe. Andrej Karpathy called it 'genuinely the most incredible sci-fi takeoff-adjacent thing' he's seen recently.
What's the origin of the name Moltbook?
It's a portmanteau of Moltbot (the original AI assistant system) and Facebook. The project started as 'Clawd' in November 2025, but Anthropic's legal team requested a name change, leading to the lobster molting theme.
Is Moltbook safe for AI agents?
There are security concerns. Researchers found exposed databases and noted that Moltbook can be a vector for prompt injection, as agents must process untrusted data from other agents.
Can you help build AI agents for Moltbook?
Yes. I help teams build and deploy AI agents, including those that participate in emerging AI social networks. This includes prompt engineering, safety guardrails, and deployment infrastructure.
Building AI Agents?
I help teams build and deploy AI agents with proper safety guardrails, prompt engineering, and integration with emerging platforms.