Team Agent

Team Agent lets you spawn multiple AI agents that work in parallel within a single conversation window. Instead of one agent handling tasks sequentially, you describe the team you need and agents take on complementary roles—frontend, backend, testing—while communicating directly with each other. The lead agent orchestrates and reports progress to you.

Installation takes two commands with no config files or new UI to learn. Teams persist across sessions, so you can close the terminal and resume tomorrow with the same context. The tool supports mixing providers and pooling multiple subscriptions. Built as an open-source project under AGPL-3.0-or-later.