AI Pulse · Under the Radar

A new tool called cumora picked up 948 stars on GitHub in its first 24 hours, which is a fast start for a small project built by one known indie developer, yetone. Worth a look, even if it stays a niche tool.

The idea is simple and a bit different from most AI chat products. Instead of bolting a chatbot onto your existing Slack, cumora builds a whole workspace where AI agents sit alongside your human team as actual teammates. You get DMs, group chats, and a shared Kanban board, but some of the people in those chats and assigned to those Kanban cards are AI agents doing real work. You can run it on cumora’s own cloud, or bring your own Claude Code or Codex subscription if you’d rather use accounts you already pay for. There’s already a live SaaS site up, so this isn’t just a code repo, it’s something you can sign up for and try today.

For a business owner, the interesting bit isn’t the stars count, it’s the design choice. Most AI tools today are still “human asks, AI answers” in a sidebar. Cumora is testing whether agents work better as teammates with their own identity, their own task cards, and their own place in the conversation history, the same way you’d track a contractor’s work. If that pattern catches on, it changes how you’d structure a small team: fewer new hires, more agents with defined roles sitting in your project boards. It’s early, the project is a day old, and there’s no track record yet on reliability or cost at scale. But this is exactly the kind of workflow pattern, agents with real accountability inside a shared workspace, that we think about when we build out an AI command centre for clients. Keep an eye on this one rather than betting on it yet. Check the repo yourself if you want to poke around before anyone else writes it up.