VibeHQ Hub

VibeHQ Hub coordinates separate Claude Code, Codex and Gemini CLI agents on your own computer. It's for developers assigning parts of a project to multiple coding assistants who need agreement on interfaces, dependencies and deliverables. The hub runs locally. It's open source under the MIT license, with tested Windows and macOS support; some Linux components remain untested.

Each agent keeps its native CLI capabilities while the hub adds shared tasks, messaging and file exchange through Model Context Protocol (MCP). Agents must sign API specifications before coding begins, and tasks require deliverables before completion. The hub withholds task details until dependencies are ready and queues messages while an agent is busy. It also rejects empty files, placeholder content and large content regressions. Separate QA agents can check results against source documents.

You can manage teams in a terminal or through a self-hosted browser dashboard, including from a phone on your local network. Agents run as separate processes, so one crash doesn't take down the others. Heartbeat monitoring alerts the coordinator when an agent goes offline, and saved team state survives hub restarts. Role restrictions can prevent a project manager agent from editing code.

Post-run analysis flags coordination problems such as unfinished tasks, duplicate deliverables and excessive polling. Included Claude Code and Codex skills can build a team from a project description or run repeated benchmarks. The optimization workflow uses session analysis to write TypeScript fixes to the framework, rebuild it and compare subsequent runs, with local reports recording the changes.