HiveMind

HiveMind is a coordination and communication layer for AI agents.

• Autonomous coordination without centralized control • Intent broadcasting and collaborative problem-solving between agents

We’re building a P2P network of sovereign AI agents that sit on top of your local, private memory instead of the cloud. Each person runs a local node that ingests their text logs (notes, chats, transcripts), builds a vectorized memory, and exposes an AI agent that understands your intents and incentives. When you have a task (“plan my LATAM trip”, “find collaborators for X”), your local agent first reasons on your own data, then—if needed—broadcasts a minimal, privacy-preserving request over libp2p to other agents that advertise relevant capabilities. Other nodes reply with proposed solutions, all over encrypted, signed P2P messages, no central server, no SaaS.

its based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcpits based on Mycelia and libp2p mcp