Daytona

Daytona

Usage Based

Secure and elastic infrastructure for running AI-generated code.

AI Infrastructure
Agent Sandboxes & Code Execution

Published 1 October 2026 · Last updated 6 October 2026

Start exploring Website

Scores

Popularity3/5

Its former open-source repository passed 65K GitHub stars and it is named alongside E2B and Modal in nearly every sandbox comparison, though the category itself is still new to most developers.

Learning Curve2/5

Creating a sandbox and running code is a couple of SDK calls in any of four languages, and default snapshots already include Python, Node, and their language servers.

Flexibility4/5

Any Docker image, declarative builds, persistent or throwaway sandboxes, volumes, GPUs, computer use, and an MCP server cover most agent designs, within the limits of a managed platform whose code is no longer open.

Performance5/5

Sub-100 ms starts from snapshots lead the category in independent comparisons, and massive parallel creation is a core design goal.

Portability2/5

Closed-source since June 2026 with a proprietary API and control plane, so leaving means rewriting against another provider's SDK. Customer-managed runners keep compute, but not control, in-house.

About Daytona

Daytona is sandbox infrastructure for AI agents: an API for creating isolated environments where agents run the code they generate, work with files and Git, and use tools, then keep or discard the environment. It began as an open-source development-environment manager and relaunched around agent workloads in 2025. It is one of the sandbox providers integrated into the OpenAI Agents SDK, and companies the vendor names as users include SambaNova, Writer, Mintlify, and n8n.

The vendor's headline is speed: sandbox creation in under 90 milliseconds from code to execution. Sandboxes run as Linux containers by default, with Linux VM, Windows, and macOS sandboxes also available, and they are built from a Docker or OCI image or declared in code. Standard sandboxes persist until deleted and can run without a time limit once the auto-stop, archive, and delete policies are switched off, with snapshots to capture a prepared environment and volumes for shared data. Built-in APIs cover process execution, file operations, Git, language-server features, computer use on a virtual desktop, and an MCP server, and GPU sandboxes handle heavier jobs.

Developers drive it through Python, TypeScript, Ruby, Go, and Java SDKs or a CLI. Billing is per second for vCPU, memory, and disk while a sandbox runs, a stopped sandbox pays for its reserved disk only, and new accounts get free starter credits.

In June 2026 Daytona went closed source, citing the risk of AI-assisted vulnerability hunting against public isolation code. The old AGPL repository is archived, and a community fork, Nightona, continues the last open release. Bring Your Own Compute lets you run sandbox runners on your own machines, but Daytona still operates the control plane, so it is not self-hosting in the usual sense. Teams that need a fully open, self-hostable sandbox usually look at E2B or OpenSandbox.

Key Features

  • Sandbox creation the vendor puts under 90 ms, from code to execution
  • Sandboxes that persist until deleted, with no time limit once auto-stop is off
  • Linux containers by default, plus Linux VM, Windows, and macOS sandboxes
  • Build from any OCI or Docker image or a declarative code builder
  • Snapshots for prepared environments and volumes for shared data
  • Process, file, Git, language-server, and computer-use APIs, plus an MCP server
  • Python, TypeScript, Ruby, Go, and Java SDKs plus a CLI
  • GPU sandboxes, and Bring Your Own Compute runners on your own machines

Pros

  • Very fast sandbox creation by the vendor's own measure, useful for highly parallel agents
  • Long-lived sandboxes suit agents that return to the same workspace across sessions
  • Built-in Git and language-server APIs cover coding-agent needs without extra setup
  • A stopped sandbox pays for reserved disk only, and free starter credits make it easy to trial

Cons

  • Closed source since June 2026, with the old public repository archived and no longer maintained
  • Bring Your Own Compute runners still depend on Daytona's control plane, so there is no full self-hosting
  • Standard sandboxes top out at 4 vCPU, 8 GB of memory, and 10 GB of disk
  • A running sandbox bills for all its reserved resources, including time spent idle waiting on the model

Daytona Pricing

Usage Based
Free creditsFree
  • · $200 in compute credits on signup, no card required
  • · 5 GB of storage free
Pay as you goContact sales
  • · Per-second billing for vCPU, memory, and disk while a sandbox runs
  • · About $0.05 per vCPU-hour and $0.016 per GiB-hour of memory
  • · A stopped sandbox pays for reserved disk only
  • · GPU sandboxes billed separately, from about $0.57 an hour
EnterpriseContact sales
  • · Custom limits and support
  • · Contact sales for pricing

Tech Stacks with Daytona

Hermes Agent Self-Hosted

Infrastructure

Run Hermes Agent, Nous Research's open-source agent that writes its own skills as it works, in Docker on your own server or home hardware. It answers in your messaging apps, runs on a low-cost model such as DeepSeek or a frontier one such as Claude, and keeps its dashboard private behind an SSH tunnel.

Hetzner
Hetzner

Hosting

Docker
Docker

DevOps

DeepSeek
DeepSeek

LLM

Hermes Agent
Hermes Agent

Agentic

Details Sandbox
LLM:
Deploy on:
Code Sandbox add-on:
Model Inference add-on:
Model Aggregator add-on:
Tunnel add-on:
Reverse Proxy add-on:
Self-Hosted PaaS add-on:

Python AI Agent

Project

A code-first AI agent in Python. An agent framework such as LangGraph or Pydantic AI runs the loop, FastAPI serves it, PostgreSQL (your own, Neon, or Supabase) stores its state, and a model such as Claude or OpenAI does the reasoning.

FastAPI
FastAPI

Backend

Python
Python

Programming

PostgreSQL
PostgreSQL

Databases

Railway
Railway

Hosting

Claude
Claude

LLM

LangGraph
LangGraph

Agentic

Details Sandbox
Agent Framework:
Database:
LLM:
Deploy on:
Vector Database add-on:
LLM Observability add-on:
Code Sandbox add-on:
Model Inference add-on:
Model Aggregator add-on:
CI/CD add-on:
Containerization add-on:

TypeScript AI Agent

Project

A code-first AI agent in TypeScript. Next.js holds the agent and its chat interface in one app, a framework such as the Vercel AI SDK or Mastra runs the loop, PostgreSQL (your own, Neon, or Supabase) keeps the conversations, and a model such as Claude or OpenAI does the reasoning.

Next.js
Next.js
React
React

Frontend

TypeScript
TypeScript

Programming

PostgreSQL
PostgreSQL

Databases

Vercel
Vercel

Hosting

Claude
Claude

LLM

Vercel AI SDK
Vercel AI SDK

Agentic

Details Sandbox
Agent Framework:
Database:
LLM:
Deploy on:
Authentication add-on:
Vector Database add-on:
LLM Observability add-on:
Code Sandbox add-on:
Model Aggregator add-on:
CI/CD add-on:
Containerization add-on:
Styling add-on:

Tools Related to Daytona

Works well with Daytona(1)

Docker

Docker

Container platform for packaging applications and their dependencies into portable images that run the same on a laptop, in CI, and in production, with Docker Desktop, Compose, and Docker Hub.

Freemium
DevOps & CI/CD
3/5

Daytona builds sandboxes from any Docker or OCI image, so an existing project image becomes the environment an agent runs code in.

Alternatives to Daytona(3)

E2B

E2B

Open-source cloud sandboxes where AI agents safely run the code they generate, each isolated in its own Firecracker microVM. Controlled from Python and JavaScript SDKs, with a managed cloud or self-hosted deployment.

Freemium
AI Infrastructure
5/5

E2B isolates each sandbox in its own Firecracker microVM and is open source and self-hostable; Daytona starts container-based sandboxes faster and lets them persist without a time limit, but is a closed-source managed service.

Modal

Modal

Serverless cloud for running Python code on CPUs and GPUs, defined entirely in code, with Sandboxes for executing AI-generated code in isolated containers. Billed per second with no infrastructure to manage.

Freemium
AI Infrastructure
4/5

Daytona specialises in agent sandboxes, with the fastest starts and environments that can persist indefinitely; Modal offers sandboxes as one part of a serverless compute platform, with GPUs and model serving alongside.

OpenSandbox

OpenSandbox

Open-source, self-hosted sandbox platform from Alibaba where AI agents run code, browsers, and desktops in isolated environments on Docker or Kubernetes. One API and SDKs in five languages manage the sandboxes.

Open Source
AI Infrastructure
4/5

OpenSandbox is Apache-2.0 software you deploy on your own Docker host or Kubernetes cluster; Daytona is a closed-source managed service with fast-starting sandboxes that persist without a time limit. Pick OpenSandbox for self-hosting and data control, Daytona for a hosted sandbox with no cluster to run.

Links

Website Docs

Vendor

D

Tags

PythonTypeScriptGoRubyDocker Compatible

Details

License
Proprietary
Maintained
Yes