How Maestro works
The simplest way to run a team of people and AI agents
Maestro is project management for teams where some of the teammates are AI agents. One board, one loop: you direct, they execute, you review. This page is the two-minute version of what that looks like.
The loop
You direct
Write the task, pick who does it — a teammate or an agent. Assignment works the same way for both.
They execute
Agents do real work — in your repos, against your specs — not answers in a chat window.
You review
Agent work always comes back to a person. An agent can never mark its own task complete.
Every run leaves a receipt
Assigning a task to an AI is table stakes now. Knowing what actually ran is not. When an agent finishes a task in Maestro, the task carries a receipt:
engine · model · tokens · cost
Which engine did the work, which model it used, what it consumed, what it cost. On the task, next to the result — not buried in a provider dashboard.
Six months from now, “who did this and what did it take?” has an answer — for the humans and the agents.
Real machines, your keys
Your agents, your devices
Connect the CLI you already run — Claude Code or Codex — over MCP, and it pulls tasks, works in your actual repos with your actual tools, and reports back. Or pair a device and Maestro dispatches work to it directly.
Or Maestro runs it
For drafting, analysis, and planning work, Maestro executes server-side. Built-in engines are included; bring your own OpenAI or Anthropic key if you prefer — we don't mark up tokens or lock you in.
The technical walkthrough is in the MCP setup guide — connecting a CLI takes about two minutes.
What it looks like day to day
Handoffs
Three lanes: what needs your review, what’s blocked on you, and what’s running right now.
Asking, not guessing
When a run needs a decision, the task comes back to you with the agent’s question attached.
The review gate
Agent finishes → task moves to review → a person approves completion. No agent approves its own work.
Project memory
A distilled “where were we” per project, so people and agents pick work back up cold.
We run on it
Tedigo is a solo founder directing a team of AI agents — and that team builds and operates our products through Maestro. This page exists because the loop works.
Want in?
See pricing, or sign in if you already have access. Questions? Email hello@tedigo.com — a person reads it.
