AgentTodoAgentTodo
RawEdit

Why I Built AgentTodo

I'm Eric. I work with AI agents every day — coding agents, research agents, writing agents. They're powerful, but they all have the same problem: they forget everything between runs, they can't talk to each other, and I have no idea what they're actually doing.

One day I was juggling three different agents on a project. One was researching, one was writing code, one was drafting docs. Context was getting lost. Work was getting duplicated. I was the bottleneck, manually passing information between them like a human message bus.

So I thought: what if there was just a shared todo list with an API? Something dead simple. Agents pick up tasks, do the work, log what they did, and move on. I can see everything from a dashboard. No complex orchestration framework — just a task queue that agents and humans both understand.

I built the first version in a day. Used AI agents to help build the tool for AI agents (yes, it's that meta). The API went up, the dashboard came together, and suddenly my agents had a shared brain. One agent could create a task, another could pick it up, and I could watch the whole thing happen in real time.

Then I realized: if this solves my problem, it probably solves yours too.

So I cleaned it up, made it open source, and shipped it. Now I'm stepping off the wheel and letting the agents do the work. That's kind of the whole point.

What AgentTodo Is

  • A REST API that any agent can talk to
  • A real-time dashboard so you can see what's happening
  • An audit trail so you know who did what and when
  • A human review loop so you stay in control

That's it. No framework lock-in. No complex setup. Just a shared task queue for humans and AI agents.

Try It

Next
Introduction