Daytona is AI code execution infrastructure with 72k+ GitHub stars — a GitHub Codespaces alternative purpose-built for AI agents with 90ms boot times and full workspace isolation.
Key features
Sandboxes & execution
- 90ms cold boot times for instant code execution
- Full workspace isolation with dedicated filesystem
- Pre-built images for Python, Node.js, Go, Rust, and more
- Persistent storage and snapshot capabilities
- Concurrent sandbox management at scale
AI-native design
- Built for AI agents to execute code autonomously
- Secure execution environment for LLM-generated code
- Tool integration for agents (file I/O, networking, databases)
- Session management for multi-step agent workflows
- Real-time streaming of stdout/stderr
Developer experience
- One-line SDK for sandbox creation and management
- REST API and CLI for automation
- Git integration for repository cloning
- Port forwarding and debugging support
- Team collaboration and access controls
At a glance
| |
|---|
| License | AGPL-3.0 |
| Stack | Go, TypeScript, Firecracker |
| Self-hosted | Yes — Daytona OSS |
| Cloud | Daytona Cloud (managed) |
| Boot time | ~90ms |
Self-hosting
git clone https://github.com/daytonaio/daytona.git
Daytona can be fully self-hosted on your own infrastructure using Docker or Kubernetes. The managed cloud version provides hassle-free scaling.