Daytona
A sandboxed code-execution platform for AI-generated code and agent workflows, providing container-based sandboxes created in under 90ms, stateful persistent workspaces, environment snapshots and optional GPU compute. Available as a managed cloud with customer-managed compute.
- Website
- https://www.daytona.io/
- Who it's for
- Platform teams running coding agents, code interpreters, evaluations and computer-use workloads that need fast, isolated, stateful sandboxes.
- Pricing
- Usage-based after $200 of free compute: vCPU $0.0504/h, memory $0.0162/GiB/h, storage $0.000108/GiB-h after 5 GiB free. GPUs metered separately.
- Category
- Agent operations
- Last verified
- 2026-10-22 (current)
- Confidence
- high
Key facts
- Managed cloud service; the hosted product is actively developed
- The public GitHub repository was archived on 3 October 2026 and is no longer maintained
- Core development moved to a private codebase as of June 2026
- Sandboxes create in under 90ms from OCI/Docker-compatible images
- Control plane is managed by Daytona; compute can run on your own cloud
- SDKs for Python, TypeScript, Go, Java and Ruby
Sources
Every claim above is drawn from these; we do not publish unsourced specifications.
- https://www.daytona.io/
- https://github.com/daytonaio/daytona
- https://www.developersdigest.tech/agent-sandboxes/daytona-oss
Notes & uncertainties
Important split: the commercial cloud is live, but the open-source repository is archived and unmaintained. If you forked it expecting upstream fixes, you will not get them. No fully self-hosted control plane is offered.
Something changed or out of date? Tell us — corrections are how this stays worth reading.