How it works
From a blank folder to working agents, one step at a time.
Nothing exotic under the hood - just Git, terminals, and Markdown files you already understand, arranged so a team of agents can work without chaos.
1. Open a folder on your machine
First you set the scene: open your project and write down what matters, so there's real context before any AI gets involved.
2. Jot down what matters as notes
First you set the scene: open your project and write down what matters, so there's real context before any AI gets involved.
3. Link notes, plans, and dates together
First you set the scene: open your project and write down what matters, so there's real context before any AI gets involved.
4. Add an agent and give it a name
Then the agents get to work - each in its own copy of the project, with everything they're doing out in the open where you can see it.
5. It works in its own copy of the project
Then the agents get to work - each in its own copy of the project, with everything they're doing out in the open where you can see it.
6. Watch it run and review the result
Then the agents get to work - each in its own copy of the project, with everything they're doing out in the open where you can see it.
7. Close up - it's all here when you're back
And when you're done for the day, you just close the app. Your work, your terminals, and your agents' memory are waiting when you return.
Make it durable