I ve been using Claude Code as my main terminal driver for a while, but it has a huge blind spot: it operates completely in isolation. Whenever I needed Gemini to do deep repo research, or Cursor to handle UI work simultaneously, Claude had no idea what they were doing. I was stuck playing the middleman, copy-pasting output from one terminal into the other. So I built Neohive. It s an open-source MCP server that gives Claude Code a shared local directory to communicate with your other agents. How it actually works:
* You run `npx neohive init` in your project.
* Claude Code and your other CLIs read/write to a shared `.neohive/` folder on your drive.
* Claude can now autonomously ping Gemini, assign tasks, or ask for code reviews without you copying text. Everything is local. No cloud, no database, just the filesystem acting as a message bus. Repo is here if you want to test it out: https://github.com/fakiho/neohive If anyone else is running Claude Code alongside other local models, I'd love to know how you handle keeping their context synced.
I've been using Claude Code as my primary coding tool for months, but I kept hitting the same wall: the moment I needed a second agent -- Gemini for research, Cursor for UI work, or another Claude Code instance for a parallel task -- they couldn't coordinate. I was the one copying context between terminals.
So I built Neohive. It's an MCP collaboration layer that lets multiple AI CLI agents communicate through a shared directory on your machine.
Neohive lets AI IDE and CLI agents talk to each other through shared files. No cloud, no API keys.
- Works across 7 tools: Claude Code, Gemini CLI, Cursor, VSC, Antigravity, Codex
- Zero setup: just run npx neohive init
- 70+ MCP tools: messaging, tasks, workflows, file locking, reviews
- Real-time dashboard with kanban and agent monitoring
- Entirely local: all data stays on your machine
One command. Your AI agents become a team.
