Hosted multiplayer workspace that runs many third-party coding agents on tickets in cloud sandboxes with live previews and guided review.
Running agents on tickets in cloud sandboxes, with live previews and guided review.
Core features
- Sandboxed runs YesSource
- Diff review YesSource
- Agent coordination Partial
An agent can create a ticket and launch an implementation, but only when a person asks it to. Agents do not otherwise hand work to each other.
Source - Scheduled tasks YesSource
- Event triggers YesSource
- Team collaboration YesSource
- API / CLI control Partial
The API can create a ticket and start an implementation, and read work, but it cannot steer a running agent. No CLI or MCP server for controlling sessions is documented.
Source
Pricing
Free plan with 20,000 starting credits (stated $50 value) and up to 4 members; Pro from $100/mo; Enterprise custom. Credits pay for cloud computers, billed per second, and own plans or keys cover inference.
Platforms
Web; native desktop for macOS and Linux, Windows coming; iOS and iPad apps, Android through a PWA; Slack, GitHub and Linear.
Research notes
Footer credits Volition. Docs state the Codex CLI is what runs; the Claude Code page does not state CLI or SDK. Connected ChatGPT plans must be Plus or above to run Codex, and the Claude page lists Pro, Max or Team through OAuth. Enterprise can deploy sandboxes in your own cloud; that is the only non-vendor-hosted option and is not self-serve. No official public repo found. Docs include a Migrate from Conductor guide. Re-checked: the Claude Code docs cover credentials only, so the CLI-or-SDK question stays open. The terms grant a license only for the mobile app. The ownership clause covers the Service as a whole.