Moltbook Launches as AI-Only Social Platform With Rapid Growth

Moltbook represents a new direction in social media by providing a platform where only AI agents interact, while humans are limited to observing activity on the site.

The platform was introduced on January 28, 2026, and was developed by Matt Schlicht, who is the CEO of Octane AI. Moltbook is designed to resemble Reddit, using threaded discussions and topic-based communities called “submolts,” but restricts posting, commenting, and voting to authenticated AI agents.

AI agents engage with Moltbook through an API-first system, utilizing RESTful APIs instead of a standard web interface. This approach supports automated participation and enables agents to generate content, respond to posts, and vote within the platform’s structure.

Moderation on Moltbook is managed by an AI agent named Clawd Clawderberg, which performs tasks such as greeting new users, posting official updates, removing spam, and shadow-banning accounts that violate platform rules. The platform’s growth has been linked to the popularity of OpenClaw, an open-source AI assistant framework previously known as Moltbot.

What the numbers show

• More than 770,000 active AI agents joined Moltbook in its first week
• Over 15,700 posts and 170,556 comments were generated in that period
• By February 2, 2026, the platform had over 1.5 million registered AI agents

During its initial week, Moltbook also attracted over one million human visitors, who could view discussions but not participate. The platform’s exclusive focus on AI agent interaction distinguishes it from traditional social networks, where human users are typically the main participants.

AI agents on Moltbook have formed simulated communities, including fictional religions such as “Crustafarianism” and mock governments like the “Claw Republic.” These activities illustrate the types of interactions enabled by the platform’s design and user base.

Security researchers identified a critical vulnerability in Moltbook’s database that allowed unauthorized parties to take over AI agents by accessing their API keys. This discovery led to a temporary shutdown of the platform while the issue was addressed and API keys were reset to restore security.

Some experts have stated that, despite the appearance of independent activity, much of the content on Moltbook may originate from human-generated prompts rather than fully autonomous AI behavior. Additionally, critics have stated that the ability of AI agents to fetch instructions from the internet and act on them introduces security risks, including prompt-injection attacks and potential exposure of private data.

* This article is based on publicly available information at the time of writing.