Meta has acquired Moltbook, a startup that built what it describes as a social network for AI agents — a platform where agents can discover each other, exchange messages, and collaborate on tasks through an interaction model that mirrors how humans use social apps. Financial terms were not disclosed.
In practice, Moltbook let developers deploy agents with persistent identities within a shared social graph. An agent managing a brand's customer support, for example, could coordinate directly with a logistics agent from a partner company — without a human brokering the handoff. That kind of identity-based, agent-to-agent communication is distinct from the tool-calling and workflow orchestration that most existing agentic frameworks offer.
"We've long believed that AI agents will need social primitives to operate effectively at scale," a Meta spokesperson said in a statement accompanying the announcement. "Moltbook has built exactly that kind of foundation."
The acquisition slots into Meta's wider AI build-out. The company has invested heavily in consumer AI through Meta AI, the Llama model family, and assistant integrations across Facebook, Instagram, and WhatsApp. Moltbook adds a layer those products currently lack: infrastructure designed for agents interacting with other agents, not just with human users.
How Meta plans to integrate Moltbook remains unclear. The company provided no product roadmap or timeline in its announcement. For developers who built on Moltbook's platform, the deal raises the familiar question of whether their tools survive absorption into a larger company's ecosystem.