daytona: a secure and elastic infrastructure runtime for executing AI-generated code in isolated sandboxes

What it solves

Daytona provides a secure, isolated, and elastic infrastructure for executing code generated by AI and running AI agent workflows. It eliminates the risks of running untrusted AI-generated code on local machines by providing dedicated, disposable sandboxes.

How it works

The platform uses "sandboxes"—fully composable computers with their own kernel, filesystem, network stack, vCPU, RAM, and disk. These sandboxes are OCI/Docker compatible and can spin up in under 90ms. Developers and AI agents interact with these environments programmatically via SDKs (Python, TypeScript, Ruby, Go, Java), a REST API, or a CLI to manage the sandbox lifecycle, execute code, and perform filesystem operations.

Who it’s for

It is designed for AI agent architects and developers who need a safe, scalable environment to run AI-generated code and maintain stateful agent operations across sessions.

Highlights

• Rapid Deployment: Sandboxes launch from code to execution in under 90ms.
• Full Isolation: Each sandbox is a complete computer with its own dedicated resources and network stack.
• State Persistence: Supports snapshots and volumes to maintain agent state across different sessions.
• Broad Integration: Provides comprehensive SDKs for multiple languages and supports the Model Context Protocol (MCP) server.