The work that held.

An open-source tool that runs every feature branch of your app at the same time on localhost, each in its own container behind a gateway. One dashboard to spin them up, tail logs, preview the running app, diff UI states, and leave QA notes per feature. It auto-detects the stack and manages multiple repos through git worktrees.

I built it because reviewing several branches at once meant constantly tearing down and rebuilding one local environment. Fleet just runs them all in parallel and lets you flip between them.

What I am proud of: it is a real tool, not a demo, and it found its people. It is niche by design, built for engineers doing parallel QA, and it has genuine adoption among them. The container orchestration, the gateway proxy, the stack auto-detection, and the diff view all had to work together, and they do.

A local AFK agent swarm for software work. You give it a goal and it builds a web of tasks, figures out the order they have to run in, and executes the whole graph autonomously: writing code, recovering from failures, and chaining the follow-up work without a human in the loop. The topology view lets you watch the dependency web plan and run itself in real time.

It is the most ambitious thing I have built. Coordinating many agents so they do the right work in the right order, and recover when a step fails, is a genuinely hard orchestration problem, and Mars does it.

Note: a private product, so there is no public link. The diagram above mirrors the live topology view — nodes are tasks, edges are blocker dependencies, colours are lifecycle state, and the failed branch shows a recovered task picked up by a follow-up agent.