Managing dozens of long-running AI sessions is becoming the choke point for developers and small teams. Instead of one-off chat windows, this project treats agents as schedulable, branchable processes you can monitor, queue, and run unattended — so complex multi-step workflows and parallel coding tasks finish without continuous manual oversight.
What Sets It Apart
- Git worktree-aware agent workspaces: create isolated sub-agents that operate in separate worktrees and produce PR-ready branches. This turns agent output into reviewable Git artifacts instead of ephemeral chat responses, reducing merge conflicts and manual reconciliation.
- Auto Run & Playbooks: file-system driven task runner that batch-processes markdown checklists through agents. Each checklist item runs in a fresh session with clean context, enabling repeatable automation loops (e.g., test/patch/evaluate) and long-running unattended runs.
- True parallel multi-agent coordination: run unlimited agents and terminal sessions in parallel, plus group chat where a moderator AI routes questions and synthesizes results. This is useful for cross-cutting architecture decisions or distributed code generation tasks.
- Keyboard-first UX + headless automation: rapid keyboard shortcuts, a CLI for cron/CI usage, and a small built-in web server for remote monitoring. The combination makes it easy to integrate into developer workflows and CI pipelines.
Who it's for & trade-offs
Great fit if you: small dev teams or solo hackers who run many agent-driven tasks in parallel; projects that want agent outputs as Git artifacts; users who prefer keyboard-driven, local-first tooling and occasional remote control.
Look elsewhere if you: need a cloud-hosted multi-tenant orchestration platform with built-in enterprise access controls out of the box, or if you require tight vendor-managed integrations for proprietary LLMs not supported by the app. The tool is a pass-through to your configured agent providers (it relies on your provider credentials and MCP tooling), so backend model capabilities and limits remain provider-dependent.
