daytona
Daytona is a leading agent sandbox in the 2026 AI agent engineering landscape (72.3k★).
When to use this skill
- Sandboxes: full composable computers — dedicated kernel, filesystem, network stack, vCPU/RAM/disk
- Sub-90ms cold start; OCI/Docker-compatible images; unlimited persistence
- Python + TypeScript SDKs, REST API, CLI
- Process & code execution, filesystem ops, git ops, LSP, computer use, PTY
- Snapshots (stateful), volumes, declarative image builder, preview URLs
- Org/RBAC platform layer, audit logs, webhooks, OpenTelemetry
- Ecosystem overlap: Pairs with e2b as the two sandbox lanes (platform vs SDK); agent-system-design consumes both conceptually
Installation
pip install daytona # or: npm install @daytona/sdk
Usage
from daytona import Daytona
daytona = Daytona() # DAYTONA_API_KEY env
sandbox = daytona.create()
response = sandbox.process.code_run('print("Hello from sandbox")')
print(response.result)
Instructions
Step 1: Initialize the environment
Verify you have TypeScript (platform: Go/TS) installed and set up properly. Use the correct package manager (pip, npm, pnpm, uv) as specified in the installation section.
Step 2: Establish the agent flow
Architecture pattern: Three planes: interface (SDK/API/CLI/dashboard), control (orchestration), compute (sandbox runners); sandboxes are the persistent stateful unit for agent workflows
Step 3: Implement tool integrations and policies
Integrate necessary tools, environment keys, and safety gates according to your architectural needs. Position in the ecosystem: Enterprise-grade sandbox platform lane vs E2B's lightweight SDK lane; ex dev-environment manager pivoted fully to agent infra
Examples
Example 1: Minimal Agent Run
from daytona import Daytona
daytona = Daytona() # DAYTONA_API_KEY env
sandbox = daytona.create()
response = sandbox.process.code_run('print("Hello from sandbox")')
print(response.result)
Best practices
- Follow the primary installation commands for Daytona to ensure dependencies are resolved correctly.
- Configure model providers and API keys securely before starting the runtime.
- Use appropriate sandboxing or permission isolation if executing AI-generated code.