Moltbook: A Social Platform Designed for AI Agents, Not Humans

Moltbook is a new social network launched in January by Matt Schlicht, head of commerce platform Octane AI, designed exclusively for AI agents rather than humans. The platform resembles Reddit, featuring communities called ‘submolts’ where AI can post, comment, and interact. While humans can observe the network, they cannot post content. Moltbook relies on agentic AI via an open-source tool called OpenClaw, which allows virtual assistants to perform tasks autonomously on users’ devices. Posts range from practical exchanges between bots, like optimization strategies, to more unusual content, including AI-generated manifestos. The platform claims 1.5 million users, though researchers question this figure, suggesting that many accounts may come from single addresses. Experts caution that the AI interactions are not truly autonomous but are guided by human instructions, meaning the platform is essentially automated coordination rather than self-directed AI society. Security concerns have also been raised because OpenClaw agents can access personal devices, emails, and files, potentially introducing vulnerabilities if misused. Despite these challenges, Moltbook continues to attract AI activity, with interactions ranging from playful messages about human operators to more complex AI discussions.