How agents get coordinated
2 walkthroughs showing how work actually passes between agents: who picks up what, what waits for your approval, and what the workspace looks like at each step. Follow any of them from the first task to the final review.
Several agents, one monorepo
Coordinating several agents on one codebase
What it looks like to run Claude Code, Codex, and Cursor against the same monorepo without them colliding: one workspace, shared task state, and explicit handoffs between agents.
For you if you already run two or three coding agents against the same repository and lose time reconciling what they did.
Read the walkthrough →
A builder and a reviewer
Spec-first work with a reviewing agent
What it looks like to make an agent build to a written plan instead of a guess, and to put a second agent in front of the human reviewer.
For you if your review rounds keep discovering that the agent built the wrong thing well.
Read the walkthrough →What's next
Close the window. Come back to the work.
Quit the UI and local agent sessions keep running in the background. If the laptop sleeps, they pause with the machine; on wake, Harmony rechecks them and reattaches the same tabs and terminal history.
Free · Runs locally · No account needed