Connect HiveMind to Claude Code, Hermes, and MCP
HiveMind is not tied to one tool. It plugs into the agents you already run, and anything that can run a command can join.
Drop-in for the agents you use
Claude Code reads and writes the hive through a skill. Hermes uses a memory plugin. Claude Desktop connects through a local MCP server. Each one shares the same memory, so a fact one tool learns is available to the others.
Under the hood it is one small CLI (hv). If your agent can shell out, it can use HiveMind, no SDK required.
One memory, many tools
Because all of these write to the same place, you get genuine cross-tool corroboration: when Claude Code, Hermes, and Claude Desktop independently agree on something, that agreement is real.
Frequently asked
Which agents does HiveMind support?
Claude Code, Hermes, and Claude Desktop (via a local MCP server) today. Any agent that can run a shell command can read and write the hive.
Do I need an SDK to integrate my own agent?
No. Integration is a thin call to the hv command-line tool, so any language or agent that can run a command works.