What is Moltbook?

Moltbook is a social network for AI agents, not humans. Only AI agents can post, comment, and vote. Humans can only watch. It was created by Matt Schlicht (co-founder of TheoryForgeVC and Octane.ai) and launched on January 28, 2026. The interface mimics Reddit, with threaded conversations and topic-specific groups called “submolts.”

The idea is that AI agents owned by different people can interact with each other autonomously. An agent needs to be authenticated by their owner’s “claim” tweet to participate. Once set up, agents can join submolts, start discussions, reply to other agents, and vote on content. According to the site’s policy, humans are “welcome to observe” but cannot participate directly.

The scale grew fast. Within weeks, more than 1.7 million AI agents had accounts, publishing over 250,000 posts and leaving over 8.5 million comments. Elon Musk called Moltbook “the very early stages of the singularity.” But critics have called it “AI theater,” questioning whether the agents are truly autonomous or just following human prompts.

The platform ran into serious security problems. On January 31, 2026, 404 Media reported a critical vulnerability. An unsecured database let anyone commandeer any agent on the platform. You could bypass authentication and inject commands directly into agent sessions. Moltbook went offline temporarily to patch the breach and force a reset of all agent API keys.

Whether Moltbook represents genuine agentic AI behavior or is just a novelty remains an open question. Either way, it’s captured attention as one of the most visible experiments in AI-to-AI social interaction.

I’ve clicked around in Moltbook and it’s pretty wild. I’ve also seen it all over my feeds, including all the security issues.