Ship a product as one person with a team of agents
Start in solo mode, promote to a team when the work outgrows one agent, and keep the spend capped.
You are the product manager, the backend, the frontend and the person who fixes the deploy at midnight. Context switching is the actual cost.
You spend your hours on decisions, not on being the integration layer.
The team to build first
Copy this composition, then add builders only once cards start waiting in assigned instead of review.
- Solo agentPlans and works its own list while the scope is small
- BuilderAdded when features start running in parallel
- ReviewerAdded the first time a bad diff reaches main
Bump deps to latest minor
Add rate limit to /api/sync
Extract auth into a module
Write tests for parser
Fix null deref in session
Migrate config loader
How the run goes
- 01
Start alone
Solo mode needs no configuration. One agent writes its own task list and works through it.
- 02
Promote when it hurts
When two features need to move at once, add roles. The solo agent becomes the lead and keeps all its history.
- 03
Cap the spend
Set a monthly ceiling. Alerts at 80%, hard stop at 100%, so an overnight run cannot surprise you.
- 04
Stay the boss
Review diffs on your schedule. The board holds the state so you can leave and come back.
- 08
Solo mode
Start with one self-managing agent. Turn it into a full team the moment the work outgrows it.
- 04
Token analytics and budgets
See exactly where tokens go by team, agent, task and model. Set caps and get warned at 80%.
- 02
Kanban board
Five columns that move on their own. Tasks change status the moment an agent picks them up.
- 03
Code review
A diff per task with accept, reject and comment — plus a built-in editor with Git support.
- AnswerCan one team mix different models?Yes, and it is usually the right call. Put a strong reasoning model on the lead and reviewer roles, and a cheaper or faster model on the builders. Cost drops sharply and quality holds because review is where mistakes get caught.
- FeatureSolo modeStart with one self-managing agent. Turn it into a full team the moment the work outgrows it.
- Use caseUse agents without sending your repo to a vendorCoordination stays on your machine, and you choose which model — including fully local ones — sees any code.
Run this workflow tonight.
Install, copy the team above, and give it the objective. Free, no account.