Mco

Most AI coding tools push you into their own model. Hive takes the opposite approach: it orchestrates whatever CLI agents you already have — Claude Code, Codex, Gemini, OpenCode, Antigravity, Cursor CLI, Grok Build, Hermes, Qwen Code — all in one browser window. Everything runs locally; no accounts, no API keys. The core workflow is refreshingly simple: an Orchestrator sends tasks via `team send`, workers complete them and `team report` back. You can chain write+review cycles across different models (they catch each other's mistakes), or set up test+fix loops that run automatically until green. The demo video on Bilibili shows a 7-minute walkthrough. Features like memory, task dependency graphs (.hive/tasks.md), and checkpoints (--resume) make complex projects manageable. Remote access via phone browser is optional, end-to-end encrypted, and off by default. Pricing is free — the tool is BSL 1.1 licensed, maintained by a single developer. Where it falls short: you must already have CLI agents installed and configured. No built-in model, no aggregator for API keys. It's designed for developers who are comfortable with terminals and already use multiple AI assistants. Compared to tools like Cursor or GitHub Copilot, Hive offers zero model integration — you bring your own. Best for teams that want to cross-check code with different models or automate multi-step review pipelines. Not for single-agent workflows or users new to AI coding assistants.