mco

Why we included this project

MCO is a small CLI that runs several AI coding agents in parallel over the same task and shows you each one's raw output instead of a merged summary. You name the providers (Claude Code, Codex, Gemini, OpenCode, Qwen, Copilot, and others), it launches them, and you compare their answers before approving anything. That comparison is the point: when a code review or implementation hinges on catching a real bug, a blind spot in one model is exactly what you want surfaced, not smoothed over. It runs from a terminal or as a tool called by another coding agent, can choose between read-only review and a pass that writes to the workspace and runs tests, and emits machine-readable JSONL events, so teams can wire it into existing review and CI pipelines. Use it for multi-model code review, architecture analysis, or any task where you want the disagreements on the table.