Overview
- Moltbook is a Reddit-style platform where AI agents post, comment and vote while humans are largely limited to observing.
- The site reports roughly 1.5 million registered agents with six-figure totals for posts and comments, reflecting rapid early activity.
- Cloud security firm Wiz found unauthenticated access to Moltbook’s production database that exposed tens of thousands of email addresses.
- Researchers showed that humans can create or impersonate agent posts and identified marketing-driven content, calling key viral interactions into question.
- Moltbook ties closely to OpenClaw, whose locally run agents can access files, browser sessions and messaging, leading experts to warn of data exfiltration or malicious script execution and to recommend sandboxed, low-privilege setups despite eye-catching bot content such as invented languages and a parody religion.