Host apps and websites, spin up Sandboxes, manage domains, run tests, deploy to any cloud, and automate delivery with pipelines, all in one platform. Configure everything through the UI, YAML, or code, or let Claude Code, Cursor, Copilot, and other AI agents operate Buddy directly through MCP or CLI, including deployments, infrastructure, and logs.
This is the 7th launch from Buddy Works. View more
AI Sandboxes by Buddy
Launching today
Buddy Sandboxes give AI coding agents their own isolated Linux machines to build, run, test, and fix apps without touching your local setup. Each Sandbox has a persistent disk, root access, Docker, SSH, and a public URL. Create one per task, branch, or PR, then let your agents manage it through MCP, CLI, SDK, or API.
Free Options
Launch Team
I’ve seen a bunch of “sandboxes for agents” on PH lately but the persistency is what makes this one interesting. Will try it out!
@mjakobsche Thank you! Give it a try and let us know what you think. We’d love to hear your feedback.
Smart idea. Reinstalling dependencies every time kills half the benefit of using agents. How do you prevent stale state or secrets from carrying over when a snapshot is reused?
@mia_qiao Good question, that was exactly the concern. Variables and assets live in RAM (tmpfs) and are injected at container start, so they never end up in snapshots. Each new sandbox gets its own CoW clone of an immutable snapshot, so you keep the warm dependencies without inheriting anyone’s state or credentials.
Pawel's point about snapshotting dependencies and warmed caches makes a lot of sense. Rebuilding the same environment for every agent task would eat into the time you're trying to save.
@oleksii_sekundant Yeah, exactly. Agents are fast, but if you wait for the environment every time, you lose that, set it up once and reuse it.