Three-Layer Multi-Agent Architecture
A manager agent guards user intent, execution agents drive the work, and SubAgents dive into sub-tasks.
An observable team of agents — from understanding the codebase to editing, testing, and review — that carries long tasks through to the end.

Not one-off chat completions — parallel, observable, and long-running.
A manager agent guards user intent, execution agents drive the work, and SubAgents dive into sub-tasks.
Tool calls and SubAgent tasks run in parallel — multi-module investigations finish markedly faster than serial chat loops.
While the main conversation keeps moving, every sub-task reports its state and results. Parallelism is no longer hidden.
Large tool outputs are cached and retrieved on demand, long histories compact automatically — keeping long sessions stable.
Local projects, remote hosts, and cloud containers under one roof — workspace-level isolation with a consistent workflow.
Code intelligence, terminal, Git, MCP tools, and Skills all flow through one execution loop, from analysis to delivery.
Multiple SubAgents push tasks forward at once, with status and results shown in the same workspace.

From tutorials to downloads — a few steps into parallel workflows.
Up and running in 10 minutes: connect your workspace, learn the UI, ship your first task.
Complete reference for architecture, configuration, API, and self-hosted deployment.
Desktop and mobile apps for macOS / Windows / Linux.
What changed and what shipped in every release.
Download the desktop app, or try it right in the browser.