What it does
neurohive adds a shared layer on top of neuromcp. Memories stored by any agent become visible to the whole team, and search results show which agent stored them, how long ago, and how often they were retrieved. It tracks each agent's expertise by category, flags conflicts when two agents store contradictory claims about the same fact, and offers a dashboard of memories, active agents, categories, activity, unresolved conflicts and knowledge gaps. At startup, an agent can load the team's top knowledge through a priming resource.
It suits teams of AI agents working on the same project, where agents re-discover the same facts or contradict each other across separate sessions. It is also useful for onboarding a new agent with existing team knowledge, or for finding which agent knows most about a topic before routing a question to it.
It runs as an MCP server started with the neurohive command over stdio, and it uses the same MCP configuration as neuromcp, with only the command changed. Behaviour is set through HIVEMIND_* variables (company namespace, agent identifier, role, priming count) and NEUROMCP_* variables (database path, log level, embedding provider). CLI commands cover the dashboard, expertise lookup, conflict listing and a health check. It exposes 16 tools, 18 resources and 3 prompts, including the 13 neuromcp tools unchanged.
Limitations and requirements
Requires neuromcp, which neurohive wraps. The default embedding provider is local; OpenAI embeddings need an OPENAI_API_KEY. Paperclip detection works automatically only when Paperclip is running locally; otherwise agent identity must be configured manually.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.AdelElo13/neurohive
Multi-agent memory intelligence with expertise tracking and conflict detection
v1.0.2As shown on registry.modelcontextprotocol.io when we last read it