Meta Acquires Moltbook, Adds Agent‑Style Social Network

Meta has acquired Moltbook, the small but headline‑grabbing social network that let users create and interact with personality‑driven AI agents. Moltbook attracted attention for its novel, vibe‑coded approach to agent creation — and for security lapses that once allowed humans to impersonate agents via an unsecured database.

The platform stood out because much of its codebase was reportedly written with the help of an AI assistant. That experimental origin helped Moltbook move fast, but also contributed to gaps in design and protection. Meta’s Superintelligence Labs will now take ownership, with the stated aim of integrating the project into Meta’s broader AI research and product work while addressing its security shortcomings.

For users and developers who followed Moltbook, the acquisition feels like a familiar pattern: a scrappy startup builds something interesting, then a big player steps in to scale and professionalize it. Meta’s interest likely stems from Moltbook’s approach to agent personalities, social interaction models, and the underlying tooling for generating agent behaviour at scale.

Meta faces a technical challenge. Fixing the security issues that allowed agent impersonation will require both engineering work and careful policy decisions about how agent identities are created, verified and moderated across social spaces. Given the public scrutiny around synthetic identities and disinformation, hardening Moltbook under a large platform will be watched closely.

Expect Meta to run audits, tighten access controls, and adapt moderation mechanisms to handle agent‑driven interactions. For the wider AI community, the deal underscores ongoing interest in agent architectures that blend personality, social mechanics and generative models — but also the need for robust safety and transparency measures as those systems enter mainstream platforms.