Conductor is a tool for running coding agents in the cloud, with the core goal of executing multiple coding agents in parallel in isolated cloud sandboxes and supporting real-time multiplayer collaboration.
Purpose Conductor is aimed at developers or teams that need to run multiple coding agents at the same time, letting agents work continuously in independent cloud environments without occupying local resources.
Main capabilities
- Cloud sandboxes: each agent has its own independent cloud sandbox, running on isolated microVMs, preloaded with the user's repositories and dependencies; sandboxes can start in seconds, and agents can run for hours.
- Multi-agent parallelism: supports running a group of coding agents in the cloud, described officially as "running a team of coding agents."
- Built-in agents: runs first-party Claude Code, Codex, Cursor, and OpenCode agents under the hood.
- Bring your own subscriptions and keys: users can use their own subscriptions and API keys.
- Multiplayer collaboration: Conductor supports multiplayer mode, allowing users to share a workspace link with teammates via a keyboard shortcut (⌘⇧C), see who is active, follow work they are interested in, and jointly instruct agents in real time.
- Multi-platform access: can be operated from the desktop app, mobile, or your own integrations via the Conductor API. Conductor for iOS is also officially available.
How to use Users can use Conductor through the desktop app, mobile, or their own integrations connected to the Conductor API. The company mentions that Conductor for iOS supports signing in with ChatGPT. The platform runs built-in Claude Code, Codex, Cursor, and OpenCode agents, and users can bring their own subscriptions and keys. For collaboration, users invite teammates by sharing a workspace link, view status in real time, and jointly prompt agents.
Use cases
- Running multiple coding agents in parallel in the cloud, suitable for developers who need to handle multiple coding tasks at the same time.
- Team collaborative development, where multiple people can share a workspace and collaborate in real time to prompt agents.
- Scenarios that require long-running agent tasks, where sandboxes can stay running for hours.
- Users who want to integrate agents into their own workflows via the API.
Limitations The materials do not provide specific pricing, free quota, supported operating systems, programming languages, or specific performance metrics. Some features (such as the detailed capabilities of the Conductor API and the specific requirements of the iOS app) are not yet described in detail.
