--- description: Skill learned from https://e2b.dev/docs — 60 pages scraped tags: learned, docs, e2b source: https://e2b.dev/docs --- # e2b > Learned from [https://e2b.dev/docs](https://e2b.dev/docs) — 60 pages ## Documentation - E2B > Source: https://e2b.dev/docs Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Getting started E2B Documentation ​ What is E2B? E2B provides isolated sandboxes that let agents safely execute code, process data, and run tools. Our SDKs make it easy to start and manage these environments. Start a sandbox and run code in a few lines: JavaScript & TypeScript Python Copy Ask AI npm i e2b JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create () // Needs E2B_API_KEY environment variable const result = await sandbox.commands. run ( 'echo "Hello from E2B Sandbox!"' ) console. log (result.stdout) ​ E2B building blocks A quick overview of the core building blocks you’ll interact with when using E2B. Sandbox — A fast, secure Linux VM created on demand for your agent Template — Defines what environment a sandbox starts with ​ How to use the docs The documentation is split into three main sections: Quickstart — Step-by-step tutorials that walk you through creating your first E2B sandboxes. Examples — In-depth tutorials focused on specific use cases. Pick the topics that match what you’re building. SDK Reference — A complete technical reference for every SDK method, parameter, and configuration option. ​ Quickstart ​ Examples Computer Use Build AI agents that see, understand, and control virtual Linux desktops using E2B Desktop sandboxes. GitHub Actions CI/CD Use E2B sandboxes in your GitHub Actions workflows to run testing, validation, and AI code reviews. Was this page helpful? Yes No Suggest edits Raise issue Running your first Sandbox This guide will show you how to start your first E2B Sandbox. Next ⌘ I On this page What is E2B? E2B building blocks How to use the docs Quickstart Examples Assistant Responses are generated using AI and may contain mistakes. --- ## SDK Reference - E2B > Source: https://e2b.dev/docs/sdk-reference SDK Reference - E2B Documentation Documentation SDK Reference This is the SDK reference for the E2B CLI & Core SDK , Code Interpreter SDK , and Desktop Sandbox SDK . The reference is intended to be limited to low-level descriptions of various programmatic functionality. If you’re just getting started with E2B, we would instead recommend looking at the documentation first or to get started quickly with an example. --- ## Documentation - E2B > Source: https://e2b.dev/docs/api-key Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Getting started API key To use the API key, you can either: Set the API key as the E2B_API_KEY environment variable to avoid passing it each time you create a sandbox. Or pass it directly to the Sandbox constructor as shown below: JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create ({ apiKey: 'YOUR_API_KEY' }) ​ Where to find API key You can get your API key at dashboard . ​ Access token The access token is used only in the CLI and is not needed in the SDK . There’s no need to set it when logging into the CLI using e2b auth login . To authenticate without the browser, you can set E2B_ACCESS_TOKEN as an environment variable. This is useful for CI/CD pipelines. ​ Where to find access token You can get your Access token key at the dashboard . Was this page helpful? Yes No Suggest edits Raise issue Previous Cookbook Next ⌘ I On this page Where to find API key Access token Where to find access token Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/support Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Getting started Need help? Join our Discord community, open GitHub issues, or send us an email if you have any questions or request. Join our Discord community Visit our GitHub Contact us Was this page helpful? Yes No Suggest edits Raise issue Previous Billing Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/billing Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Getting started Billing & pricing E2B uses usage-based pricing - you pay only for what you use. New users receive $100 in free credits to get started. Manage billing in dashboard ​ Plans Feature Hobby Pro Enterprise Base price $0/month $150/month Custom Free credits $100 (one-time) $100 (one-time) Custom Max session length 1 hour 24 hours Custom Concurrent sandboxes 20 100 - 1,100 1,100+ Pro plan includes 100 concurrent sandboxes. Higher concurrency up to 1,100 is available as a purchasable add-on . Plans have different API rate limits . To upgrade your plan or purchase add-ons, visit the dashboard billing tab . For Enterprise plans, contact sales . ​ Usage-based pricing You pay per second for compute resources while your sandbox is running. ​ Compute costs Use the usage cost calculator on our pricing page to estimate costs for your specific configuration. ​ Customizing compute resources You can customize allocated CPU and RAM when building custom templates by specifying cpuCount and memoryMB in the build configuration. JavaScript & TypeScript Python Copy Ask AI import { Template, defaultBuildLogger } from 'e2b' await Template. build (template, 'my-template' , { cpuCount: 8 , memoryMB: 4096 , onBuildLogs: defaultBuildLogger (), }) See template quickstart for more details on building custom templates. Need higher CPU or RAM limits? Contact support@e2b.dev for more information. ​ Monitoring usage Check your usage and costs in the dashboard usage tab . ​ FAQ When am I charged? Automatically at the start of the month for the previous month’s usage. What happens when I run out of credits? Your account will be blocked. Add a payment method to continue using E2B. Can I set spending limits? Yes, you can set spending limits on the budget page in your dashboard. How do I optimize costs? Enable auto-pause - Automatically pause sandboxes after a period of inactivity to stop billing while preserving state Pause sandboxes when idle - Use sbx.pause() to stop billing while keeping state available for later Kill sandboxes you no longer need - Use sbx.kill() to stop billing and release resources permanently Allocate only what you need ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/use-cases/coding-agents Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Use cases Coding Agents Coding agents like Claude Code , Codex , and Amp can write, debug, and refactor code autonomously. E2B sandboxes give each agent a full Linux environment with terminal, filesystem, and git — completely isolated from your infrastructure. Pre-built templates mean you can go from zero to a running agent in a single API call. ​ Why Use a Sandbox Running coding agents directly on your machine or servers means giving AI-generated code unrestricted access to your environment. E2B sandboxes solve this: Isolation — agent-generated code runs in a secure sandbox, never touching your production systems or local machine Full dev environment — terminal, filesystem, git, and package managers are all available out of the box, so agents work like a developer would Pre-built templates — ready-made templates for popular agents get you started fast, and you can build your own for any agent Scalability — run many sandboxes in parallel, each with its own agent on a separate task ​ How It Works Create a sandbox — use a pre-built template or build your own with any agent installed Agent gets a full environment — terminal, filesystem, git access, and any tools installed in the template Agent works autonomously — it reads the codebase, writes code, runs tests, and iterates until the task is done Extract results — pull out the git diff, structured output, or modified files via the SDK. The sandbox stays available for follow-up work, or you can pause it to pick up later ​ Agent Examples Since each sandbox is a full Linux environment, you can run any coding agent — just install it in a custom template . E2B also provides pre-built templates for popular agents to get you started quickly. Claude Code Anthropic’s autonomous coding agent with structured output and MCP tool support Codex OpenAI’s coding agent with schema-validated output and image input Amp Coding agent with streaming JSON and thread management OpenCode Open-source multi-provider agent with a built-in web UI ​ Related Guides Git Integration Clone repos, manage branches, and push changes from sandboxes Sandbox Persistence Pause and resume sandboxes to preserve state across sessions Custom Templates Build your own sandbox templates with custom tool ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/use-cases/computer-use Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Use cases Computer use Computer use agents interact with graphical desktops the same way a human would — viewing the screen, clicking, typing, and scrolling. E2B provides the sandboxed desktop environment where these agents operate safely, with VNC streaming for real-time visual feedback. For a complete working implementation, see E2B Surf — an open-source computer use agent you can try via the live demo . ​ How it works The computer use agent loop follows this pattern: User sends a command — e.g., “Open Firefox and search for AI news” Agent creates a desktop sandbox — an Ubuntu 22.04 environment with XFCE desktop and pre-installed applications Agent takes a screenshot — captures the current desktop state via E2B Desktop SDK LLM analyzes the screenshot — a vision model (e.g., OpenAI Computer Use API ) decides what action to take Action is executed — click, type, scroll, or keypress via E2B Desktop SDK Repeat — new screenshot is taken and sent back to the LLM until the task is complete ​ Install the E2B Desktop SDK The E2B Desktop SDK gives your agent a full Linux desktop with mouse, keyboard, and screen capture APIs. JavaScript & TypeScript Python Copy Ask AI npm i @e2b/desktop ​ Core implementation The following snippets are adapted from E2B Surf . ​ Setting up the sandbox Create a desktop sandbox and start VNC streaming so you can view the desktop in a browser. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/desktop' // Create a desktop sandbox with a 5-minute timeout const sandbox = await Sandbox. create ({ resolution: [ 1024 , 720 ], dpi: 96 , timeoutMs: 300_000 , }) // Start VNC streaming for browser-based viewing await sandbox.stream. start () const streamUrl = sandbox.stream. getUrl () console. log ( 'View desktop at:' , streamUrl) ​ Executing desktop actions The E2B Desktop SDK maps directly to mouse and keyboard actions. Here’s how Surf translates LLM-returned actions into desktop interactions. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/desktop' const sandbox = await Sandbox. create ({ timeoutMs: 300_000 }) // Mouse action ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/use-cases/ci-cd Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Use cases GitHub Actions CI/CD CI/CD pipelines can use AI agents to review pull requests, generate tests, and validate code changes automatically. E2B sandboxes provide the secure, isolated execution environment where these agents can safely clone repositories, run untrusted code, and report results — all triggered by GitHub Actions on every pull request. Each run uses its own isolated sandbox, so malicious or buggy PR code never touches your CI runner. ​ GitHub Actions workflow The workflow triggers on pull request events and runs a review script. E2B_API_KEY and the LLM API key are stored as GitHub Actions secrets , while the built-in GITHUB_TOKEN is available automatically. The permissions block grants write access so the script can post PR comments. .github/workflows/ai-review.yml (JavaScript) .github/workflows/ai-review.yml (Python) Copy Ask AI name : AI Code Review on : pull_request : types : [ opened , synchronize ] permissions : pull-requests : write jobs : ai-review : runs-on : ubuntu-latest steps : - uses : actions/checkout@v4 - name : Set up Node.js uses : actions/setup-node@v4 with : node-version : "20" - name : Install dependencies run : npm install e2b openai - name : Run AI review env : E2B_API_KEY : ${{ secrets.E2B_API_KEY }} OPENAI_API_KEY : ${{ secrets.OPENAI_API_KEY }} GITHUB_TOKEN : ${{ secrets.GITHUB_TOKEN }} PR_REPO : ${{ github.event.pull_request.head.repo.full_name }} PR_BRANCH : ${{ github.event.pull_request.head.ref }} PR_NUMBER : ${{ github.event.pull_request.number }} GITHUB_REPOSITORY : ${{ github.repository }} run : node review.mjs ​ Review script The workflow calls this script on every PR. It runs five steps inside an E2B sandbox, keeping all untrusted code isolated from the CI runner. review.mjs review.py Copy Ask AI import { Sandbox, CommandExitError } from 'e2b' import OpenAI from 'openai' // --- 1. Create sandbox --- const sandbox = await Sandbox. create ({ timeoutMs: 300_000 }) conso ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/agents/amp Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Agents in sandbox Amp Amp is a coding agent with multi-model architecture and built-in code intelligence. E2B provides a pre-built amp template with Amp already installed. ​ CLI Create a sandbox with the E2B CLI . Copy Ask AI e2b sbx create amp Once inside the sandbox, start Amp. Copy Ask AI amp ​ Run headless Use -x for non-interactive mode and --dangerously-allow-all to auto-approve all tool calls (safe inside E2B sandboxes). Amp uses its own API key from ampcode.com/settings . JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'amp' , { envs: { AMP_API_KEY: process.env. AMP_API_KEY }, }) const result = await sandbox.commands. run ( `amp --dangerously-allow-all -x "Create a hello world HTTP server in Go"` ) console. log (result.stdout) await sandbox. kill () ​ Example: work on a cloned repository JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'amp' , { envs: { AMP_API_KEY: process.env. AMP_API_KEY }, timeoutMs: 600_000 , }) await sandbox.git. clone ( 'https://github.com/your-org/your-repo.git' , { path: '/home/user/repo' , username: 'x-access-token' , password: process.env. GITHUB_TOKEN , depth: 1 , }) const result = await sandbox.commands. run ( `cd /home/user/repo && amp --dangerously-allow-all -x "Add error handling to all API endpoints"` , { onStdout : ( data ) => process.stdout. write (data) } ) const diff = await sandbox.commands. run ( 'cd /home/user/repo && git diff' ) console. log (diff.stdout) await sandbox. kill () ​ Streaming JSON Use --stream-json to get a real-time JSONL event stream with rich metadata — including tool calls, token usage, thinking blocks, and permission decisions. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'amp' , { envs: { AMP_API_KEY: process.env. AMP_API_KEY }, }) const result = awa ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/agents/claude-code Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Agents in sandbox Claude Code Claude Code is Anthropic’s agentic coding tool. E2B provides a pre-built claude template with Claude Code already installed. ​ CLI Create a sandbox with the E2B CLI . Copy Ask AI e2b sbx create claude Once inside the sandbox, start Claude Code. Copy Ask AI claude ​ Run headless Use -p for non-interactive mode and --dangerously-skip-permissions to auto-approve all tool calls (safe inside E2B sandboxes). JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'claude' , { envs: { ANTHROPIC_API_KEY: process.env. ANTHROPIC_API_KEY }, }) const result = await sandbox.commands. run ( `claude --dangerously-skip-permissions -p "Create a hello world HTTP server in Go"` ) console. log (result.stdout) await sandbox. kill () ​ Example: work on a cloned repository JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'claude' , { envs: { ANTHROPIC_API_KEY: process.env. ANTHROPIC_API_KEY }, timeoutMs: 600_000 , }) await sandbox.git. clone ( 'https://github.com/your-org/your-repo.git' , { path: '/home/user/repo' , username: 'x-access-token' , password: process.env. GITHUB_TOKEN , depth: 1 , }) const result = await sandbox.commands. run ( `cd /home/user/repo && claude --dangerously-skip-permissions -p "Add error handling to all API endpoints"` , { onStdout : ( data ) => process.stdout. write (data) } ) const diff = await sandbox.commands. run ( 'cd /home/user/repo && git diff' ) console. log (diff.stdout) await sandbox. kill () ​ Structured output Use --output-format json to get machine-readable responses — useful for building pipelines or extracting specific results. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'claude' , { envs: { ANTHROPIC_API_KEY: process.env. ANTHROPIC_API_KEY }, }) const result = await sandbox.c ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/agents/codex Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Agents in sandbox Codex Codex is OpenAI’s open-source coding agent. E2B provides a pre-built codex template with Codex already installed. ​ CLI Create a sandbox with the E2B CLI . Copy Ask AI e2b sbx create codex Once inside the sandbox, start Codex. Copy Ask AI codex ​ Run headless Use codex exec for non-interactive mode and --full-auto to auto-approve tool calls (safe inside E2B sandboxes). Pass --skip-git-repo-check to bypass git directory ownership checks inside the sandbox. Pass CODEX_API_KEY as an environment variable. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'codex' , { envs: { CODEX_API_KEY: process.env. CODEX_API_KEY }, }) const result = await sandbox.commands. run ( `codex exec --full-auto --skip-git-repo-check "Create a hello world HTTP server in Go"` ) console. log (result.stdout) await sandbox. kill () ​ Example: work on a cloned repository Use -C to set Codex’s working directory to a cloned repo. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'codex' , { envs: { CODEX_API_KEY: process.env. CODEX_API_KEY }, timeoutMs: 600_000 , }) await sandbox.git. clone ( 'https://github.com/your-org/your-repo.git' , { path: '/home/user/repo' , username: 'x-access-token' , password: process.env. GITHUB_TOKEN , depth: 1 , }) const result = await sandbox.commands. run ( `codex exec --full-auto --skip-git-repo-check -C /home/user/repo "Add error handling to all API endpoints"` , { onStdout : ( data ) => process.stdout. write (data) } ) const diff = await sandbox.commands. run ( 'cd /home/user/repo && git diff' ) console. log (diff.stdout) await sandbox. kill () ​ Schema-validated output Use --output-schema to constrain the agent’s final response to a JSON Schema. This ensures the output conforms to a specific structure — useful for building reliable pipelines. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } fr ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/agents/opencode Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Agents in sandbox OpenCode OpenCode is an open-source coding agent that supports multiple LLM providers. E2B provides a pre-built opencode template with OpenCode already installed. ​ CLI Create a sandbox with the E2B CLI . Copy Ask AI e2b sbx create opencode Once inside the sandbox, start OpenCode. Copy Ask AI opencode ​ Run headless Use opencode run for non-interactive mode. Pass your LLM provider’s API key as an environment variable — OpenCode supports ANTHROPIC_API_KEY , OPENAI_API_KEY , GEMINI_API_KEY , and others . JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'opencode' , { envs: { ANTHROPIC_API_KEY: process.env. ANTHROPIC_API_KEY }, }) const result = await sandbox.commands. run ( `opencode run "Create a hello world HTTP server in Go"` ) console. log (result.stdout) await sandbox. kill () ​ Example: work on a cloned repository JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( 'opencode' , { envs: { ANTHROPIC_API_KEY: process.env. ANTHROPIC_API_KEY }, timeoutMs: 600_000 , }) await sandbox.git. clone ( 'https://github.com/your-org/your-repo.git' , { path: '/home/user/repo' , username: 'x-access-token' , password: process.env. GITHUB_TOKEN , depth: 1 , }) const result = await sandbox.commands. run ( `cd /home/user/repo && opencode run "Add error handling to all API endpoints"` , { onStdout : ( data ) => process.stdout. write (data) } ) const diff = await sandbox.commands. run ( 'cd /home/user/repo && git diff' ) console. log (diff.stdout) await sandbox. kill () ​ Connect with the OpenCode SDK OpenCode includes a headless HTTP server that you can control programmatically using the @opencode-ai/sdk client. Start the server inside a sandbox, get the public URL with sandbox.getHost() , and connect from your application. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' import { createOpenco ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/code-interpreting/streaming Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Code interpreting Streaming E2B Code Interpreter SDK allows you to stream the output, and results when executing code in the sandbox. ​ Stream stdout and stderr When using the runCode() method in JavaScript or run_code() in Python you can pass onStdout / on_stdout and onStderr / on_stderr callbacks to handle the output. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const codeToRun = ` import time import sys print("This goes first to stdout") time.sleep(3) print("This goes later to stderr", file=sys.stderr) time.sleep(3) print("This goes last") ` const sandbox = await Sandbox. create () sandbox. runCode (codeToRun, { // Use `onError` to handle runtime code errors onError : error => console. error ( 'error:' , error), onStdout : data => console. log ( 'stdout:' , data), onStderr : data => console. error ( 'stderr:' , data), }) The code above will print the following: JavaScript & TypeScript Python Copy Ask AI stdout : { error : false , line : "This goes first to stdout \n " , timestamp : 1729049666861000 , } stderr : { error : true , line : "This goes later to stderr \n " , timestamp : 1729049669924000 , } stdout : { error : false , line : "This goes last \n " , timestamp : 1729049672664000 , } ​ Stream results When using the runCode() method in JavaScript or run_code() in Python you can pass onResults / on_results callback to receive results from the sandbox like charts, tables, text, and more. JavaScript & TypeScript Python Copy Ask AI const codeToRun = ` import matplotlib.pyplot as plt # Prepare data categories = ['Category A', 'Category B', 'Category C', 'Category D'] values = [10, 20, 15, 25] # Create and customize the bar chart plt.figure(figsize=(10, 6)) plt.bar(categories, values, color='green') plt.xlabel('Categories') plt.ylabel('Values') plt.title('Values by Category') # Display the chart plt.show() ` const sandbox = await Sandbox. create () ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/code-interpreting/contexts Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Code interpreting Code contexts Run code in different code execution contexts with E2B Code Interpreter SDK. This allows you to parallelize code execution by running code in different contexts at the same time. By default the code is run in the Sandbox’s default code execution context. You can change it by passing the context parameter to the runCode() method in JavaScript or run_code() in Python. ​ Create a new Code Context You can create a new code execution context by calling the createCodeContext() method in JavaScript or create_code_context() in Python and passing the context parameters. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const context = await sandbox. createCodeContext ({ cwd: '/home/user' , language: 'python' , requestTimeoutMs: 60_000 , }) const result = await sandbox. runCode ( 'print("Hello, world!")' , { context }) console. log (result) ​ List active Code Contexts You can list active code execution contexts by calling the listCodeContexts() method in JavaScript or list_code_contexts() in Python. This will return a list of active code execution contexts. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const contexts = await sandbox. listCodeContexts () console. log (contexts) ​ Restart a Code Context You can restart an active code execution context by calling the restartCodeContext() method in JavaScript or restart_code_context() in Python and passing the context object or context ID. Restarting a context will clear its state and start a new code execution session in the same context. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const context = await sandbox. createCodeContext ({ cwd: '/home/user' , language: 'python' , requestTimeoutMs: 60_000 , }) // using context object await sandbox. restartCodeContext (contex ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox lifecycle Sandboxes stay running as long as you need them. When their timeout expires, they automatically pause to save resources — preserving their full state so you can resume at any time. You can also configure an explicit timeout or shut down a sandbox manually. Sandboxes can run continuously for up to 24 hours (Pro) or 1 hour (Base). For longer workloads, use pause and resume — pausing resets the runtime window, and your sandbox’s full state is preserved indefinitely. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create sandbox with and keep it running for 60 seconds. // 🚨 Note: The units are milliseconds. const sandbox = await Sandbox. create ({ timeoutMs: 60_000 , }) ​ Change sandbox timeout during runtime You can change the sandbox timeout when it’s running by calling the setTimeout method in JavaScript or set_timeout method in Python. When you call the set timeout method, the sandbox timeout will be reset to the new value that you specified. This can be useful if you want to extend the sandbox lifetime when it’s already running. You can for example start with a sandbox with 1 minute timeout and then periodically call set timeout every time user interacts with it in your app. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create sandbox with and keep it running for 60 seconds. const sandbox = await Sandbox. create ({ timeoutMs: 60_000 }) // Change the sandbox timeout to 30 seconds. // 🚨 The new timeout will be 30 seconds from now. await sandbox. setTimeout ( 30_000 ) ​ Retrieve sandbox information You can retrieve sandbox information like sandbox ID, template, metadata, started at/end at date by calling the getInfo method in JavaScript or get_info method in Python. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create sandbox with and keep it running for 60 seconds. const sandbox = await Sandbox. create ({ timeoutMs: 60_000 }) // Retrieve sandbox information. const info = await sandbox. getInfo () console. log (info) // { // " ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/lifecycle-events-api Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Monitor sandbox lifecycle events The lifecycle API provides RESTful endpoints to request the latest sandbox lifecycle events. This allows you to track when sandboxes are created, paused, resumed, updated, snapshotted, or killed, along with metadata. All requests require authentication using your team API key . Query Parameters: offset (optional): Number of events to skip (default: 0, min: 0) limit (optional): Number of events to return (default: 10, min: 1, max: 100) orderAsc (optional): Sort order - true for ascending, false for descending (default: false) JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sbx = await Sandbox. create () // Get the latest events for a specific sandbox const resp1 = await fetch ( `https://api.e2b.app/events/sandboxes/${ sbx . id }` , { method: 'GET' , headers: { 'X-API-Key' : E2B_API_KEY , }, } ) const sandboxEvents = await resp1. json () // Get the latest 10 events for all sandboxes associated with the team const resp2 = await fetch ( 'https://api.e2b.app/events/sandboxes?limit=10' , { method: 'GET' , headers: { 'X-API-Key' : E2B_API_KEY , }, } ) const teamSandboxEvents = await resp2. json () console. log (teamSandboxEvents) // [ // { // "version": "v1", // "id": "f5911677-cb60-498f-afed-f68143b3cc59", // "type": "sandbox.lifecycle.killed", // "eventData": null, // "sandboxBuildId": "a979a14b-bdcc-49e6-bc04-1189fc9fe7c2", // "sandboxExecutionId": "1dae9e1c-9957-4ce7-a236-a99d5779aadf", // "sandboxId": "${SANDBOX_ID}", // "sandboxTeamId": "460355b3-4f64-48f9-9a16-4442817f79f5", // "sandboxTemplateId": "rki5dems9wqfm4r03t7g", // "timestamp": "2025-08-06T20:59:36Z" // }, // { // "version": "v1", // "id": "30b09e11-9ba2-42db-9cf6-d21f0f43a234", // "type": "sandbox.lifecycle.updated", // "eventData": { // "set_timeout": "2025-08-06T20:59:59Z" // }, // "sandboxBuildId": "a979a14b-bdcc-49e6-bc04-1189fc9fe7c2", // "sandboxExecutionId": "1dae9e1c-9957-4ce7- ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/lifecycle-events-webhooks Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox lifecycle webhooks Webhooks provide a way for notifications to be delivered to an external web server whenever certain sandbox lifecycle events occur. This allows you to receive real-time updates about sandbox creation, updates, and termination without having to poll the API. All webhook requests require authentication using your team API key . ​ Webhook management ​ Register webhook Register a new webhook to receive sandbox lifecycle events. The webhook will be registered for the team ID associated with your API key. All events specified during webhook creation will be sent to URL provided during registration with following payload . JavaScript & TypeScript Python Copy Ask AI // Register a new webhook const resp = await fetch ( 'https://api.e2b.app/events/webhooks' , { method: 'POST' , headers: { 'X-API-Key' : E2B_API_KEY , 'Content-Type' : 'application/json' , }, body: JSON . stringify ({ name: 'My Sandbox Webhook' , url: 'https://your-webhook-endpoint.com/webhook' , enabled: true , events: [ 'sandbox.lifecycle.created' , 'sandbox.lifecycle.updated' , 'sandbox.lifecycle.killed' ], signatureSecret: 'secret-for-event-signature-verification' }), } ) if (resp.status === 201 ) { console. log ( 'Webhook registered successfully' ) } ​ List webhooks List all registered webhooks for your team. JavaScript & TypeScript Python Copy Ask AI // List webhooks const resp = await fetch ( 'https://api.e2b.app/events/webhooks' , { method: 'GET' , headers: { 'X-API-Key' : E2B_API_KEY }, }, ) if (resp.status === 200 ) { console. log ( 'Webhooks listed successfully' ) console. log ( await resp. json ()) } ​ Get webhook configuration Retrieve the current webhook configuration for your team. JavaScript & TypeScript Python Copy Ask AI // Get webhook configuration const resp = await fetch ( `https://api.e2b.app/events/webhooks/${ webhookID }` ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/persistence Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox persistence The sandbox persistence allows you to pause your sandbox and resume it later from the same state it was in when you paused it. This includes not only state of the sandbox’s filesystem but also the sandbox’s memory. This means all running processes, loaded variables, data, etc. ​ Sandbox state transitions Understanding how sandboxes transition between different states is crucial for managing their lifecycle effectively. Here’s a diagram showing the possible state transitions: ​ State descriptions Running : The sandbox is actively running and can execute code. This is the initial state after creation. Paused : The sandbox execution is suspended but its state is preserved. Snapshotting : The sandbox is briefly paused while a persistent snapshot is being created. It automatically returns to Running. See Snapshots . Killed : The sandbox is terminated and all resources are released. This is a terminal state. ​ Changing sandbox’s state JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Starts in Running state // Pause the sandbox await sandbox. betaPause () // Running → Paused // Resume the sandbox await sandbox. connect () // Running/Paused → Running // Kill the sandbox (from any state) await sandbox. kill () // Running/Paused → Killed ​ Pausing sandbox When you pause a sandbox, both the sandbox’s filesystem and memory state will be saved. This includes all the files in the sandbox’s filesystem and all the running processes, loaded variables, data, etc. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sbx = await Sandbox. create () console. log ( 'Sandbox created' , sbx.sandboxId) // Pause the sandbox // You can save the sandbox ID in your database to resume the sandbox later await sbx. betaPause () console. log ( 'Sandbox paused' , sbx.sandboxId) ​ Resuming sandbox When you resume a sandbox, it will be in the same state it was in when you paused it. This means that all the files in the sandbox’s filesystem will be restored and all the running processes, lo ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/snapshots Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox snapshots Snapshots let you create a persistent point-in-time capture of a running sandbox, including both its filesystem and memory state. You can then use a snapshot to spawn new sandboxes that start from the exact same state. The original sandbox continues running after the snapshot is created, and a single snapshot can be used to create many new sandboxes. ​ Snapshots vs. Pause/Resume Pause/Resume Snapshots Effect on original sandbox Pauses (stops) the sandbox Sandbox briefly pauses, then continues running Relationship One-to-one — resume restores the same sandbox One-to-many — snapshot can spawn many new sandboxes Use case Suspend and resume a single sandbox Create a reusable checkpoint For pause/resume functionality, see Persistence . ​ Snapshot flow The sandbox is briefly paused during the snapshot process but automatically returns to running state. The sandbox ID stays the same after the snapshot completes. During the snapshot, the sandbox is temporarily paused and resumed. This causes all active connections (e.g. WebSocket, PTY, command streams) to be dropped. Make sure your client handles reconnection properly. ​ Create a snapshot You can create a snapshot from a running sandbox instance. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create () // Create a snapshot from a running sandbox const snapshot = await sandbox. createSnapshot () console. log ( 'Snapshot ID:' , snapshot.snapshotId) You can also create a snapshot by sandbox ID using the static method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' // Create a snapshot by sandbox ID const snapshot = await Sandbox. createSnapshot (sandboxId) console. log ( 'Snapshot ID:' , snapshot.snapshotId) ​ Create a sandbox from a snapshot The snapshot ID can be used directly with Sandbox.create() to spawn a new sandbox from the snapshot. The new sandbox starts with the exact filesystem and memory state captured in the snapshot. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const snapshot = await sandbox. create ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/git-integration Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Git integration Use the sandbox.git methods to run common git operations inside a sandbox. ​ Authentication and Identity ​ Passing credentials inline For private repositories over HTTP(S), pass username and password (token) directly to commands that need authentication. A username is required whenever you pass a password/token. JavaScript & TypeScript Python Copy Ask AI const repoPath = '/home/user/repo' await sandbox.git. push (repoPath, { username: process.env. GIT_USERNAME , password: process.env. GIT_TOKEN , }) await sandbox.git. pull (repoPath, { username: process.env. GIT_USERNAME , password: process.env. GIT_TOKEN , }) ​ Credential helper (authenticate once) To avoid passing credentials on each command, store them in the git credential helper inside the sandbox using dangerouslyAuthenticate() / dangerously_authenticate() . Stores credentials on disk inside the sandbox. Any process or agent with access to the sandbox can read them. Use only when you understand the risk. JavaScript & TypeScript Python Copy Ask AI // Default (GitHub) await sandbox.git. dangerouslyAuthenticate ({ username: process.env. GIT_USERNAME , password: process.env. GIT_TOKEN , }) // Custom host (self-hosted) await sandbox.git. dangerouslyAuthenticate ({ username: process.env. GIT_USERNAME , password: process.env. GIT_TOKEN , host: 'git.example.com' , protocol: 'https' , }) // After this, HTTPS git operations use the stored credentials await sandbox.git. clone ( 'https://git.example.com/org/repo.git' , { path: '/home/user/repo' }) await sandbox.git. push ( '/home/user/repo' ) ​ Keep credentials in the remote URL By default, credentials are stripped from the remote URL after cloning. To keep credentials in the remote URL (stored in .git/config ), set dangerouslyStoreCredentials / dangerously_store_credentials . Storing credentials in the remote URL persists them in the repo config. Any process or agent with access to the sandbox can read them. Only use this when required. JavaScript & TypeScript Python Copy Ask AI // Default: credentials are stripped from the remote UR ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/metrics Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox metrics The sandbox metrics allows you to get information about the sandbox’s CPU, memory and disk usage. ​ Getting sandbox metrics Getting the metrics of a sandbox returns an array of timestamped metrics containing CPU, memory and disk usage information. The metrics are collected every 5 seconds. ​ Getting sandbox metrics using the SDKs JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sbx = await Sandbox. create () console. log ( 'Sandbox created' , sbx.sandboxId) // Wait for a few seconds to collect some metrics await new Promise (( resolve ) => setTimeout (resolve, 10_000 )) const metrics = await sbx. getMetrics () // You can also get the metrics by sandbox ID: // const metrics = await Sandbox.getMetrics(sbx.sandboxId) console. log ( 'Sandbox metrics:' , metrics) // Sandbox metrics: // [ // { // timestamp: 2025-07-28T08:04:05.000Z, // cpuUsedPct: 20.33, // cpuCount: 2, // memUsed: 32681984, // in bytes // memTotal: 507592704, // in bytes // diskUsed: 1514856448, // in bytes // diskTotal: 2573185024 // in bytes // }, // { // timestamp: 2025-07-28T08:04:10.000Z, // cpuUsedPct: 0.2, // cpuCount: 2, // memUsed: 33316864, // in bytes // memTotal: 507592704, // in bytes // diskUsed: 1514856448, // in bytes // diskTotal: 2573185024 // in bytes // } // ] ​ Getting sandbox metrics using the CLI Terminal Copy Ask AI e2b sandbox metrics < sandbox_i d > # Metrics for sandbox # # [2025-07-25 14:05:55Z] CPU: 8.27% / 2 Cores | Memory: 31 / 484 MiB | Disk: 1445 / 2453 MiB # [2025-07-25 14:06:00Z] CPU: 0.5% / 2 Cores | Memory: 32 / 484 MiB | Disk: 1445 / 2453 MiB # [2025-07-25 14:06:05Z] CPU: 0.1% / 2 Cores | Memory: 32 / 484 MiB | Disk: 1445 / 2453 MiB # [2025-07-25 14:06:10Z] CPU: 0.3% / 2 Cores | Memory: 32 / 484 MiB | Disk: 1445 / 2453 MiB It may take a second or more to get the first metrics after the sandbox is created. Until the first metrics are collected from the sandbox, you will get an empty array. Was this p ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/metadata Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Sandbox metadata Metadata is a way to attach arbitrary key-value pairs for a sandbox. This is useful in various scenarios, for example: Associate a sandbox with a user session. Store custom user data for a sandbox like API keys. Associate a sandbox with a user ID and connect to it later . You specify metadata when creating a sandbox and can access it later through listing running sandboxes with Sandbox.list() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create sandbox with metadata. const sandbox = await Sandbox. create ({ metadata: { userId: '123' , }, }) // List running sandboxes and access metadata. const paginator = await Sandbox. list () const runningSandboxes = await paginator. nextItems () // Will print: // { // 'userId': '123', // } console. log (runningSandboxes[ 0 ].metadata) ​ Filtering sandboxes by metadata You can also filter sandboxes by metadata, you can find more about it here . Was this page helpful? Yes No Suggest edits Raise issue Previous Environment variables Next ⌘ I On this page Filtering sandboxes by metadata Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/environment-variables Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Environment variables This page covers how to set and use environment variables in a sandbox, and default environment variables inside the sandbox. ​ Default environment variables ​ Knowing if you are inside a sandbox Sometimes it’s useful to know if the code is running inside a sandbox. Upon creating a sandbox, useful sandbox metadata is set as environment variables for commands: E2B_SANDBOX is set to true for processes to know if they are inside our VM. E2B_SANDBOX_ID to know the ID of the sandbox. E2B_TEAM_ID to know the team ID that created the sandbox. E2B_TEMPLATE_ID to know what template was used for the current sandbox. You can try it out by running the following code in the sandbox: JavaScript & TypeScript Python Copy Ask AI const sandbox = await Sandbox. create () const result = await sandbox.commands. run ( 'echo $E2B_SANDBOX_ID' ) These default environment variables are only accessible via the SDK, when using the CLI you can find them in the form of dot files in the /run/e2b/ dir: Copy Ask AI user@e2b:~$ ls -a /run/e2b/ .E2B_SANDBOX .E2B_SANDBOX_ID .E2B_TEAM_ID .E2B_TEMPLATE_ID ​ Setting environment variables There are 3 ways to set environment variables in a sandbox: Global environment variables when creating the sandbox . When running code in the sandbox . When running commands in the sandbox . ​ 1. Global environment variables You can set global environment variables when creating a sandbox. JavaScript & TypeScript Python Copy Ask AI const sandbox = await Sandbox. create ({ envs: { MY_VAR: 'my_value' , }, }) ​ 2. Setting environment variables when running code You can set environment variables for specific code execution call in the sandbox. These environment variables are scoped to the command but are not private in the OS. If you had a global environment variable with the same name, it will be overridden only for the command. JavaScript & TypeScript Python Copy Ask AI const sandbox = await Sandbox. create () const result = await sandbox. runCode ( 'import os; print(os.environ.get("MY_VAR"))' , { envs: { MY_VAR: 'my_value' , }, }) ​ 3. ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/list Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox List sandboxes You can list sandboxes using the Sandbox.list() method. Once you have information about running sandbox, you can connect to it using the Sandbox.connect() method. ​ Listing sandboxes The Sandbox.list() method supports pagination. In the advanced pagination section, you can find more information about pagination techniques using the updated method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox, SandboxInfo } from '@e2b/code-interpreter' const sandbox = await Sandbox. create ( { metadata: { name: 'My Sandbox' , }, }, ) const paginator = Sandbox. list () // Get the first page of sandboxes (running and paused) const firstPage = await paginator. nextItems () const runningSandbox = firstPage[ 0 ] console. log ( 'Running sandbox metadata:' , runningSandbox.metadata) console. log ( 'Running sandbox id:' , runningSandbox.sandboxId) console. log ( 'Running sandbox started at:' , runningSandbox.startedAt) console. log ( 'Running sandbox template id:' , runningSandbox.templateId) // Get the next page of sandboxes const nextPage = await paginator. nextItems () The code above will output something like this: JavaScript & TypeScript Python Copy Ask AI Running sandbox metadata: {name: "My Sandbox"} Running sandbox id: ixjj3iankaishgcge4jwn-b0b684e9 Running sandbox started at: 2024-10-15T21:13:07.311Z Running sandbox template id: 3e4rngfa34txe0gxc1zf ​ Filtering sandboxes Filter sandboxes by their current state. The state parameter can contain either “ running ” for running sandboxes or “ paused ” for paused sandboxes, or both. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create a sandbox. const sandbox = await Sandbox. create () // List sandboxes that are running or paused. const paginator = Sandbox. list ({ query: { state: [ 'running' , 'paused' ], }, }) const sandboxes = await paginator. nextItems () Filter sandboxes by the metadata key value pairs specified during Sa ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/connect Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Connect to running sandbox If you have a running sandbox, you can connect to it using the Sandbox.connect() method and then start controlling it with our SDK. This is useful if you want to, for example, reuse the same sandbox instance for the same user after a short period of inactivity. ​ 1. Get the sandbox ID To connect to a running sandbox, you first need to retrieve its ID. You can do this by calling the Sandbox.list() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sbx = await Sandbox. create () // Get all running sandboxes const paginator = await Sandbox. list ({ query: { state: [ 'running' ] }, }) const runningSandboxes = await paginator. nextItems () if (runningSandboxes. length === 0 ) { throw new Error ( 'No running sandboxes found' ) } // Get the ID of the sandbox you want to connect to const sandboxId = runningSandboxes[ 0 ].sandboxId ​ 2. Connect to the sandbox Now that you have the sandbox ID, you can connect to the sandbox using the Sandbox.connect() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. connect (sandboxId) // Now you can use the sandbox as usual // ... const result = await sandbox.commands. run ( "whoami" ) console. log ( `Running in sandbox ${ sandbox . sandboxId } as "${ result . stdout . trim () }"` ) Was this page helpful? Yes No Suggest edits Raise issue Previous Internet access Next ⌘ I On this page 1. Get the sandbox ID 2. Connect to the sandbox Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/internet-access Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Internet access Every sandbox has access to the internet and can be reached by a public URL. ​ Controlling internet access You can control whether a sandbox has access to the internet by using the allowInternetAccess parameter when creating a sandbox. By default, internet access is enabled ( true ), but you can disable it for security-sensitive workloads. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Create sandbox with internet access enabled (default) const sandbox = await Sandbox. create ({ allowInternetAccess: true }) // Create sandbox without internet access const isolatedSandbox = await Sandbox. create ({ allowInternetAccess: false }) When internet access is disabled, the sandbox cannot make outbound network connections, which provides an additional layer of security for sensitive code execution. Setting allowInternetAccess to false is equivalent to setting network.denyOut to ['0.0.0.0/0'] (denying all traffic). ​ Fine-grained network control For more granular control over network access, you can use the network configuration option to specify allow and deny lists for outbound traffic. ​ Allow and deny lists You can specify IP addresses, CIDR blocks, or domain names that the sandbox is allowed to use: JavaScript & TypeScript Python Copy Ask AI import { Sandbox, ALL_TRAFFIC } from '@e2b/code-interpreter' // Deny all traffic except specific IPs const sandbox = await Sandbox. create ({ network: { denyOut: [ ALL_TRAFFIC ], allowOut: [ '1.1.1.1' , '8.8.8.0/24' ] } }) // Deny specific IPs only const restrictedSandbox = await Sandbox. create ({ network: { denyOut: [ '8.8.8.8' ] } }) ​ Domain-based filtering You can allow traffic to specific domains by specifying hostnames in allow out . When using domain-based filtering, you must include ALL_TRAFFIC in deny out to block all other traffic. Domains are not supported in the deny out list. JavaScript & TypeScript Python Copy Ask AI import { Sandbox, ALL_TRAFFIC } from '@e2b/code-interpreter' // Allow only traffic to google.c ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/pty Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Interactive terminal (PTY) The PTY (pseudo-terminal) module allows you to create interactive terminal sessions in the sandbox with real-time, bidirectional communication. Unlike commands.run() which executes a command and returns output after completion, PTY provides: Real-time streaming - Output is streamed as it happens via callbacks Bidirectional input - Send input while the terminal is running Interactive shell - Full terminal support with ANSI colors and escape sequences Session persistence - Disconnect and reconnect to running sessions ​ Create a PTY session Use sandbox.pty.create() to start an interactive bash shell. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const terminal = await sandbox.pty. create ({ cols: 80 , // Terminal width in characters rows: 24 , // Terminal height in characters onData : ( data ) => { // Called whenever terminal outputs data process.stdout. write (data) }, envs: { MY_VAR: 'hello' }, // Optional environment variables cwd: '/home/user' , // Optional working directory user: 'root' , // Optional user to run as }) // terminal.pid contains the process ID console. log ( 'Terminal PID:' , terminal.pid) The PTY runs an interactive bash shell with TERM=xterm-256color , which supports ANSI colors and escape sequences. ​ Timeout PTY sessions have a configurable timeout that controls the session duration. The default is 60 seconds. For interactive or long-running sessions, set timeoutMs: 0 (JavaScript) or timeout=0 (Python) to keep the session open indefinitely. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const terminal = await sandbox.pty. create ({ cols: 80 , rows: 24 , onData : ( data ) => process.stdout. write (data), timeoutMs: 0 , // Keep the session open indefinitely }) ​ Send input to PTY Use sendInput() in JavaScript or send_stdin( ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/ssh-access Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox SSH access SSH access enables remote terminal sessions, SCP/SFTP file transfers, and integration with tools that expect SSH connectivity. ​ Quickstart 1 Build the SSH template Define a template with OpenSSH server and websocat : JavaScript & TypeScript Python Copy Ask AI // template.ts import { Template, waitForPort } from 'e2b' export const template = Template () . fromUbuntuImage ( '25.04' ) . aptInstall ([ 'openssh-server' ]) . runCmd ([ 'curl -fsSL -o /usr/local/bin/websocat https://github.com/vi/websocat/releases/latest/download/websocat.x86_64-unknown-linux-musl' , 'chmod a+x /usr/local/bin/websocat' , ], { user: 'root' }) . setStartCmd ( 'sudo websocat -b --exit-on-eof ws-l:0.0.0.0:8081 tcp:127.0.0.1:22' , waitForPort ( 8081 )) Build the template: JavaScript & TypeScript Python Copy Ask AI // build.ts import { Template, defaultBuildLogger } from 'e2b' import { template as sshTemplate } from './template' await Template. build (sshTemplate, 'ssh-ready' , { cpuCount: 2 , memoryMB: 2048 , onBuildLogs: defaultBuildLogger (), }) 2 Create a sandbox from the template JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sbx = await Sandbox. create ( 'ssh-ready' ) console. log (sbx.sandboxId) 3 Connect from your machine macOS Linux Copy Ask AI # Install websocat brew install websocat # Connect to your sandbox ssh -o 'ProxyCommand=websocat --binary -B 65536 - wss://8081-%h.e2b.app' user@ < sandbox-i d > ​ How it works This method uses websocat to proxy SSH connections over WebSocket through the sandbox’s exposed ports. Copy Ask AI ┌───────────────────────────────────────────────────────────┐ │ Your Machine │ │ ┌──────────┐ ProxyCommand ┌──────────────────┐ │ │ │ SSH │ ────────────────── │ websocat │ │ │ │ Client │ │ (WebSocket) │ │ │ └──────────┘ └─────────┬────────┘ │ └───── ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/connect-bucket Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Connecting storage bucket to the sandbox To connect a bucket for storing data from the sandbox, we will use the FUSE file system to mount the bucket to the sandbox. You will need to create a custom sandbox template with the FUSE file system installed. The guide for building a custom sandbox template can be found here . ​ Google Cloud Storage ​ Prerequisites To use Google Cloud Storage, you’ll need a bucket and a service account. You can create a service account here and a bucket here . If you want to write to the bucket, make sure the service account has the Storage Object User role for this bucket. You can find a guide on creating a service account key here . ​ Mounting the bucket To use the Google Cloud Storage we need to install the gcsfuse package. There’s simple template that can be used to create a container with the gcsfuse installed. JavaScript & TypeScript Python Copy Ask AI const template = Template () . fromTemplate ( "code-interpreter-v1" ) . aptInstall ([ "gnupg" , "lsb-release" ]) . runCmd ( "lsb_release -c -s > /tmp/lsb_release" ) . runCmd ( 'GCSFUSE_REPO=$(cat /tmp/lsb_release) && echo "deb [signed-by=/usr/share/keyrings/cloud.google.asc] https://packages.cloud.google.com/apt gcsfuse-$GCSFUSE_REPO main" | sudo tee /etc/apt/sources.list.d/gcsfuse.list' , ) . runCmd ( "curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo tee /usr/share/keyrings/cloud.google.asc" , ) . aptInstall ([ "gcsfuse" ]) The bucket is mounted during the sandbox runtime using the gcsfuse command. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' const sandbox = await Sandbox. create ( '' ) await sandbox.files. makeDir ( '/home/user/bucket' ) await sandbox.files. write ( 'key.json' , '' ) await sandbox.commands. run ( 'sudo gcsfuse --key-file /home/user/key.json /home/user/bucket' ) ​ Flags The complete list of flags is available here . ​ Allow the default user to access the files To allow the default user to access the files, we can use t ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/rate-limits Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Rate limits This page describes API and sandbox rate limits of the E2B platform. ​ Table of Contents Sandbox lifecycle and management API Sandbox operations and requests to sandboxes Concurrent Sandboxes Sandbox creation rate Egress connections Reaching limits Increasing and avoiding rate limits Here’s a limits breakdown table based on the plan: Plan Hobby Pro Enterprise Sandbox lifecycle & management API 20,000 / 30s 20,000 / 30s Custom Sandbox operations 40,000 / 60s per IP 40,000 / 60s per IP Custom Concurrent sandboxes 20 100 - 1,100 * Custom Sandbox creation rate 1 / sec 5 / sec Custom Egress connections per sandbox 2,500 2,500 Custom * Pro plan default is 100 concurrent sandboxes. Higher concurrency of up to 1,100 is available as a separate purchasable add-on. ​ Sandbox lifecycle and management API 20,000 requests per 30 seconds This rate limit applies to lifecycle and management operations such as sandbox create, kill, update, list, and other. ​ Sandbox operations and requests to sandboxes 40,000 requests per 60 seconds This rate limits applies to operations to/within running sandboxes such as running code, listing files, running commands, etc. This also includes requests made to custom ports in the sandbox. This limit is enforced globally across all sandbox operations from a single IP address. ​ Concurrent sandboxes Number of concurrent sandboxes differs based on the pricing tier. ​ Hobby tier Up to 20 concurrent sandboxes ​ Pro tier Starts at 100 concurrent sandboxes Can go up to 1,100 with additional add-ons available for purchase . ​ Enterprise tier Custom limit - 1,100+ concurrent sandboxes ​ Sandbox creation rate This limit controls how quickly you can create new sandboxes. ​ Hobby tier 1 sandbox per second ​ Pro tier 5 sandboxes per second ​ Enterprise tier Custom limit - 5+ sandboxes per second ​ Egress connections 2,500 connections per sandbox This limit controls the maximum number of outbound (egress) network connections that can be established from a single sandbox. ​ Reaching limits When you reach the limits of your plan, subsequent requests/function calls will be effectively dropped and return the following: the 429 Too Many Requests HTTP st ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/secured-access Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Secured access Secure access authenticates communication between SDK and sandbox controller. Sandbox controller runs in sandbox itself and exposes APIs for work with file system, run commands, and generally control the sandbox via our SDK. Without secure access, anyone with a sandbox ID can access the controller APIs and control the sandbox from inside. SDKs version v2.0.0 and above are using secure access by default when creating sandbox. This may not be compatible with older custom templates and you may need to rebuild them. ​ Migration path When using custom templates created before envd v0.2.0 , you need to rebuild the templates to enable secure access. Temporarily, you can disable secure access by setting secure to false during sandbox creation, but this is not recommended for production use because it increases security risks. You can check the template envd version using the e2b template list command or by viewing the templates list on the dashboard. ​ Supported versions All sandboxes based on templates with envd version at least v0.2.0 already support secure access without any additional changes. The secure access flag was introduced in 1.5.0 for JavaScript and Python SDKs to be used optionally. Starting with SDK version v2.0.0 , sandboxes are created with secure access enabled by default. ​ Access sandbox API directly In some cases, you might want to access sandbox controller APIs directly through its URL, such as when you are not using SDKs. When secure access is enabled, you must provide an authentication token that was returned during sandbox creation. Each call to the sandbox controller must include an additional header X-Access-Token with the access token value returned during sandbox creation. For sandbox upload and download URLs, you need to generate pre-signed URLs. We are advising to use SDK for generating presigned URLs. ​ Disable secure access Disabling secured access is discouraged because it creates security vulnerabilities. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create ({ secure: false }) // Explicitly disable Was this page helpful? Yes No Suggest edits Raise ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/ip-tunneling Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Proxy tunneling ​ Setting up proxy tunneling We will set up a proxy server on a GCP VM instance running Shadowsocks and use it to tunnel the Sandbox network traffic. This will allow you to use a dedicated IP address for outgoing requests. ​ GCP VM setup Create a firewall rule to allow all tcp/udp traffic to port 8388. Replace your-project-id with your actual project ID. Copy Ask AI gcloud compute firewall-rules create shadowsocks \ --project=your-project-id \ --direction=INGRESS \ --priority=1000 \ --network=default \ --action=ALLOW \ --rules=tcp:8388,udp:8388 \ --source-ranges=0.0.0.0/0 \ --target-tags=allow-shadowsocks Create a VM instance with the following tags: allow-shadowsocks . Replace your-project-id with your actual project ID. Copy Ask AI gcloud compute instances create shadowsocks-vm \ --project=your-project-id \ --zone=us-west1-a \ --machine-type=n2-standard-2 \ --can-ip-forward \ --tags=allow-shadowsocks \ --image-project=debian-cloud \ --image-family=debian-12 \ --boot-disk-size=20GB After the VM is created, you can connect to it using the following command: Replace your-project-id with your actual project ID. Copy Ask AI # May take a few seconds until the instance is ready to accept SSH connections gcloud compute ssh shadowsocks-vm \ --project=your-project-id \ --zone=us-west1-a ​ Shadowsocks server setup (VM) SSH into the VM and follow the instructions below to install and configure Shadowsocks. Install the necessary packages, if missing: Copy Ask AI sudo apt update sudo apt install -y wget tar Download and install Shadowsocks (v1.24.0): Copy Ask AI wget https://github.com/shadowsocks/shadowsocks-rust/releases/latest/download/shadowsocks-v1.24.0.x86_64-unknown-linux-gnu.tar.xz tar -xf shadowsocks- * .tar.xz sudo mv ssserver ssmanager ssurl /usr/local/bin/ Create a shadowsocks configuration file: Replace STRONG_PASSWORD_HERE with your own password. /etc/shadowsocks/server.json Copy Ask AI { "server" : "0.0.0.0" , "server_port" : 8388 , "password" : "STRONG_PASSWORD_HERE" , ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/sandbox/custom-domain Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Sandbox Custom domain How to set up a custom domain for Sandboxes hosted on E2B. We will set up a GCP VM running Caddy server with Docker and Cloudflare DNS to proxy the requests to the Sandboxes. Example: 8080-sandboxid.mydomain.com -> 8080-sandboxid.e2b.app ​ Prerequisites Domain name registered and configured with Cloudflare DNS. Cloudflare API Token that allows you to manage DNS records. ​ GCP VM setup Create a VM instance by running the following command: Replace your-project-id with your actual project ID. Copy Ask AI gcloud compute instances create e2b-custom-domain-proxy \ --project=your-project-id \ --zone=us-west1-a \ --machine-type=n2-standard-2 \ --can-ip-forward \ --tags=http-server,https-server \ --image-project=debian-cloud \ --image-family=debian-12 \ --boot-disk-size=20GB After the VM is created, you can connect to it using the following command: Replace your-project-id with your actual project ID. Copy Ask AI # May take a few seconds until the instance is ready to accept SSH connections gcloud compute ssh e2b-custom-domain-proxy \ --project=your-project-id \ --zone=us-west1-a ​ Server setup Install the latest stable version of Docker: Copy Ask AI curl -fsSL https://get.docker.com | sudo sh Create a Dockerfile that will be used to build the Caddy server image with Cloudflare DNS: Dockerfile Copy Ask AI FROM caddy:builder AS builder RUN xcaddy build \ --with github.com/caddy-dns/cloudflare FROM caddy:latest COPY --from=builder /usr/bin/caddy /usr/bin/caddy Create a Docker Compose file that will be used to start the Caddy server: docker-compose.yml Copy Ask AI services : caddy : build : context : . dockerfile : Dockerfile container_name : caddy-proxy restart : unless-stopped ports : - "80:80" - "443:443" - "443:443/udp" # Optional: HTTP/3 volumes : - ./Caddyfile:/etc/caddy/Caddyfile - caddy_data:/data - caddy_config:/config environment : - CLOUDFLARE_API_TOKEN=${CLOUDFLARE_API ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/quickstart Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Quickstart E2B templates allow you to define custom sandboxes. You can define the base image, environment variables, files to copy, commands to run, and a start command that runs during the template build and is captured in a snapshot — so the process is already running when you create a sandbox from that template. This gives you fully configured sandboxes with running processes ready to use with zero wait time for your users. There are two ways how you can start creating a new template: using the CLI manually using the SDK ​ CLI You can use the E2B CLI to create a new template. 1 Install the E2B CLI Install the latest version of the E2B CLI 2 Initialize a new template Copy Ask AI e2b template init 3 Follow the prompts Follow the prompts to create a new template. 4 Done Check the generated README.md file to see how to build and use your new template. ​ Manual ​ Install the packages Requires the E2B SDK version at least 2.3.0 JavaScript & TypeScript Python Copy Ask AI npm install e2b dotenv Create the .env file Copy Ask AI E2B_API_KEY = e2b_*** ​ Create a new template file Create a template file with the following name and content JavaScript & TypeScript Python Copy Ask AI // template.ts import { Template, waitForTimeout } from 'e2b' ; export const template = Template () . fromBaseImage () . setEnvs ({ HELLO: "Hello, World!" , }) . setStartCmd ( "echo $HELLO" , waitForTimeout ( 5_000 )); ​ Create a development build script JavaScript & TypeScript Python Copy Ask AI // build.dev.ts import 'dotenv/config' ; import { Template, defaultBuildLogger } from 'e2b' ; import { template } from './template' ; async function main () { await Template. build (template, 'template-tag-dev' , { cpuCount: 1 , memoryMB: 1024 , onBuildLogs: defaultBuildLogger (), }); } main (). catch (console.error); ​ Create a production build script JavaScript & TypeScript Python Copy Ask AI // build.prod.ts import 'dotenv/config' ; import { Template, defaultBuildLogger } from 'e2b' ; import { template ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/how-it-works Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates How it works ​ General overview Every time you are building a sandbox template, we create a container based on the definition. We extract the container’s filesystem, do provisioning and configuration (e.g. installing required dependencies), run layers commands and start a sandbox. Then, these steps happen: We take the running sandbox. (Only if you specified the start command , otherwise this step is skipped) Execute the start command. Wait for readiness (by default 20 seconds if start command is specified, otherwise immediately ready). Readiness check can be configured using ready command . Snapshot the sandbox and make it ready for you to spawn it with the SDK. We call this sandbox snapshot a sandbox template . Snapshots are saved running sandboxes. We serialize and save the whole sandbox’s filesystem together with all the running processes in a way that can be loaded later. This allows us to load the sandbox in ~80ms any time later with all the processes already running and the filesystem exactly as it was. ​ Default user and workdir To learn more about default user and workdir, please refer to the User and workdir section. ​ Caching To learn more about caching, please refer to the Caching section. ​ Kernel E2B sandboxes run on an LTS 6.1 Linux kernel . The exact kernel version available in your sandbox depends on when your template was built : Template build date (DD.MM.YYYY) Kernel version ≥ 27.11.2025 6.1.158 (current) < 27.11.2025 6.1.102 Note: Kernel versions are fixed at template build time . If your template was created on an older kernel, the kernel cannot be upgraded or changed for that template. To use a newer kernel, you must rebuild the template or create a new one. You can find the kernel configuration files for each version here . Was this page helpful? Yes No Suggest edits Raise issue Previous User and workdir Default user and working directory in the sandbox and template Next ⌘ I On this page General overview Default user and workdir Caching Kernel Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/user-and-workdir Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates User and workdir The default user in the template is user with the /home/user (home directory) as the working directory. This is different from the Docker defaults, where the default user is root with / as the working directory. This is to help with tools installation, and to improve default security. The last set user and workdir in the template is then persisted as a default to the sandbox execution. Example of setting user and workdir in the template definition are below. Requires the E2B SDK version at least 2.3.0 ​ Default user and workdir in sandbox JavaScript & TypeScript Python Copy Ask AI const sbx = await Sandbox. create () await sbx.commands. run ( "whoami" ) // user await sbx.commands. run ( "pwd" ) // /home/user ​ Custom user and workdir template JavaScript & TypeScript Python Copy Ask AI // template.ts const template = Template () . fromBaseImage () . runCmd ( "whoami" ) // user . runCmd ( "pwd" ) // /home/user . setUser ( "guest" ) . runCmd ( "whoami" ) // guest . runCmd ( "pwd" ) // /home/guest // build_dev.ts await Template. build (template, 'custom-user-template' , { onBuildLogs: defaultBuildLogger () }) // index.ts const sbx = await Sandbox. create ( "custom-user-template" ) await sbx.commands. run ( "whoami" ) // guest await sbx.commands. run ( "pwd" ) // /home/guest Was this page helpful? Yes No Suggest edits Raise issue Previous Caching How the caching process works Next ⌘ I On this page Default user and workdir in sandbox Custom user and workdir template Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/caching Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Caching The caching concept is similar to Docker’s layer caching . For each layer command ( .copy() , .runCmd() , .setEnvs() , etc.), we create a new layer on top of the existing. Each layer is cached based on the command and its inputs (e.g., files copied, command executed, environment variables set). If a layer command is unchanged and its inputs are the same as in any previous build, we reuse the cached layer instead of rebuilding it. This significantly speeds up the build process, especially for large templates with many layers. The cache is scoped to the team, so even if you have multiple templates, they can share the same cache if they have identical layers. ​ Invalidation You can invalidate the caches only partially, or for the whole template. ​ Partial Force rebuild from the next instruction, use the method: JavaScript & TypeScript Python Copy Ask AI const template = Template () . fromBaseImage () . skipCache () . runCmd ( "echo 'Hello, World!'" ) This will force rebuild from the next instruction, invalidating the cache for all subsequent instructions in the template. ​ Whole template To force rebuild the whole template, you can use also skipCache / skip_cache parameter in the Template.build method: JavaScript & TypeScript Python Copy Ask AI Template. build (template, 'my-template' , { skipCache: true , // Configure cache skip (except for files) }) This will skip the cache usage for the whole template build. ​ Files caching When using the .copy() command, we cache the files based on their content. If the files haven’t changed since the last build, we reuse them from the files cache. We differ from Docker here. Because we build the template on our infrastructure, we use improved caching on files level. Even if you invalidate the layer before .copy() (e.g., by changing ENV variables), we’ll reuse the already uploaded files. The copy() command will still be re-executed, but the files for the layer will be reused from the files cache, no need to upload them from your computer again. To invalidate the cache for all subsequent instructions in the template AND the layer files cache, use the forceUpload / force_upload parameter. JavaScript ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/base-image Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Base image ​ Creating template When creating a template, you can specify options: JavaScript & TypeScript Python Copy Ask AI const template = Template ({ fileContextPath: "." , // Custom file context path fileIgnorePatterns: [ ".git" , "node_modules" ], // File patterns to ignore }); File ignoring : The SDK automatically reads .dockerignore files and combines them with your fileIgnorePatterns (TypeScript) or file_ignore_patterns (Python). Files matching these patterns are excluded from uploads and hash calculations. ​ Defining base image Every template starts with a base image that provides the foundation for your sandbox environment. ​ Predefined base images Use convenience methods for common base images with Ubuntu, Debian, Python, Node.js, or Bun: JavaScript & TypeScript Python Copy Ask AI template. fromUbuntuImage ( "22.04" ); // ubuntu:22.04 template. fromDebianImage ( "stable-slim" ); // debian:stable-slim template. fromPythonImage ( "3.13" ); // python:3.13 template. fromNodeImage ( "lts" ); // node:lts template. fromBunImage ( "1.3" ); // oven/bun:1.3 ​ Custom base image Use any Docker image from Docker Hub or other registries: JavaScript & TypeScript Python Copy Ask AI template. fromImage ( "custom-image:latest" ); ​ Default E2B base image Use the default E2B base image, which comes pre-configured for sandbox environments: JavaScript & TypeScript Python Copy Ask AI template. fromBaseImage (); // e2bdev/base ​ Build from existing template Extend an existing template from your team or organization: JavaScript & TypeScript Python Copy Ask AI template. fromTemplate ( "my-template" ); // Your team's template template. fromTemplate ( "acme/other-template" ); // Full namespaced reference You can only call base image methods once per template. Subsequent calls will throw an error. ​ Parsing existing Dockerfiles Convert existing Dockerfiles to template format using fromDockerfile() : JavaScript & TypeScript Python Copy Ask AI const dockerfileContent = ` FROM ubuntu:22.04 RUN apt-get update && apt-get install -y curl WORKDIR /app COPY . . ENV NODE_ENV=production ENV PORT=3000 USER appuse ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/private-registries Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Private registries If your base image is hosted in a private registry, you can provide credentials using the following helpers: General registry GCP Artifact Registry AWS ECR ​ General registry JavaScript & TypeScript Python Copy Ask AI Template (). fromImage ( 'ubuntu:22.04' , { username: 'user' , password: 'pass' , }) ​ GCP Artifact Registry JavaScript & TypeScript Python Copy Ask AI // From file path Template (). fromGCPRegistry ( 'ubuntu:22.04' , { serviceAccountJSON: './service_account.json' , }) // From object Template (). fromGCPRegistry ( 'ubuntu:22.04' , { serviceAccountJSON: { project_id: '123' , private_key_id: '456' }, }) ​ AWS ECR JavaScript & TypeScript Python Copy Ask AI Template (). fromAWSRegistry ( 'ubuntu:22.04' , { accessKeyId: '123' , secretAccessKey: '456' , region: 'us-west-1' , }) Was this page helpful? Yes No Suggest edits Raise issue Previous Defining template How to create your own template Next ⌘ I On this page General registry GCP Artifact Registry AWS ECR Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/defining-template Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Defining template ​ Method chaining All template methods return the template instance, allowing for fluent API usage: JavaScript & TypeScript Python Copy Ask AI const template = Template () . fromUbuntuImage ( "22.04" ) . aptInstall ([ "curl" ]) // necessary for waitForPort . setWorkdir ( '/app' ) . copy ( "package.json" , "/app/package.json" ) . runCmd ( "npm install" ) . setStartCmd ( "npm start" , waitForPort ( 3000 )); ​ User and workdir Set the working directory and user for the template: JavaScript & TypeScript Python Copy Ask AI // Set working directory template. setWorkdir ( "/app" ); // Set user (runs subsequent commands as this user) template. setUser ( 'node' ) template. setUser ( "1000:1000" ); // User ID and group ID ​ Copying files Copy files from your local filesystem to the template: JavaScript & TypeScript Python Copy Ask AI // Copy a single file template. copy ( "package.json" , "/app/package.json" ); // Copy multiple files to same destination template. copy ([ "file1" , "file2" ], "/app/file" ) // Multiple copy operations using copyItems template. copyItems ([ { src: "src/" , dest: "/app/src/" }, { src: "package.json" , dest: "/app/package.json" }, ]); // Copy with user and mode options template. copy ( "config.json" , "/app/config.json" , { user: "appuser" , mode: 0o644 , }); ​ File operations Perform various file operations during template build: JavaScript & TypeScript Python Copy Ask AI // Remove files or directories template. remove ( "/tmp/temp-file.txt" ); template. remove ( "/old-directory" , { recursive: true }); template. remove ( "/file.txt" , { force: true }); // Rename files or directories template. rename ( "/old-name.txt" , "/new-name.txt" ); template. rename ( "/old-dir" , "/new-dir" , { force: true }); // Create directories template. makeDir ( "/app/logs" ); template. makeDir ( "/app/data" , { mode: 0o755 }); // Create symbolic links template. makeSymlink ( "/app/data" , "/app/logs/data" ); ​ Installing packages Install packages using package ma ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/start-ready-command Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Start & ready commands ​ Start command The start command lets you specify a process that runs at the end of the template build — not when a sandbox is created. During the build, E2B executes the start command, waits for the ready command to confirm the process is up, and then takes a snapshot of the entire sandbox including the running process. When you later create a sandbox from that template, the snapshotted process is already running — there is no startup wait. This is how you get servers, seeded databases, or any long-running process available instantly when spawning sandboxes with the SDK. The start command runs once during template build and is captured in a snapshot. It does not re-execute each time you create a sandbox. If you need to run a command every time a sandbox is created, use sandbox.commands.run() after creating the sandbox instead. This also means that environment variables passed to Sandbox.create() are not available to the start command process — it already ran during the build. If your start command needs environment variables, set them in the template definition using setEnvs() / set_envs() . You can see the full build process here . ​ Ready command The ready command allows you to specify a command that will determine template sandbox readiness before a snapshot is created. It is executed in an infinite loop until it returns a successful exit code 0 . This way you can control how long should we wait for the start command or any system state. ​ Usage Set the start command (executed during template build) and the ready command (determines when the process is up before snapshotting): JavaScript & TypeScript Python Copy Ask AI // Set both start command and ready command template. setStartCmd ( 'npm start' , waitForPort ( 3000 )) // Set custom start and ready command template. setStartCmd ( 'npm start' , 'curl -s -o /dev/null -w "200"' ) // Set only ready command template. setReadyCmd ( waitForTimeout ( 10_000 )) The ready command is used to determine when the sandbox is ready to accept connections. ​ Ready command helpers The SDK provides helper functions for common ready command patterns: JavaScript & TypeScript P ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/build Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Build ​ Build and wait for completion The build method builds the template and waits for the build to complete. It returns build information including the template ID and build ID. JavaScript & TypeScript Python Copy Ask AI const buildInfo = await Template. build (template, 'my-template' , { cpuCount: 2 , // CPU cores memoryMB: 2048 , // Memory in MB skipCache: false , // Configure cache skip (except for files) onBuildLogs: defaultBuildLogger (), // Log callback receives LogEntry objects apiKey: 'your-api-key' , // Override API key domain: 'your-domain' , // Override domain }) // buildInfo contains: { name, templateId, buildId } ​ Build in background The buildInBackground method starts the build process and returns immediately without waiting for completion. This is useful when you want to trigger a build and check its status later. JavaScript & TypeScript Python Copy Ask AI const buildInfo = await Template. buildInBackground (template, 'my-template' , { cpuCount: 2 , memoryMB: 2048 , }) // Returns immediately with: { name, templateId, buildId } ​ Check build status Use getBuildStatus to check the status of a build started with buildInBackground . JavaScript & TypeScript Python Copy Ask AI const status = await Template. getBuildStatus (buildInfo, { logsOffset: 0 , // Optional: offset for fetching logs }) // status contains: { status: 'building' | 'ready' | 'error', logEntries: [...] } ​ Example: Background build with status polling JavaScript & TypeScript Python Copy Ask AI // Start build in background const buildInfo = await Template. buildInBackground (template, 'my-template' , { cpuCount: 2 , memoryMB: 2048 , }) // Poll for build status let logsOffset = 0 let status = 'building' while (status === 'building' ) { const buildStatus = await Template. getBuildStatus (buildInfo, { logsOffset, }) logsOffset += buildStatus.logEntries. length status = buildStatus.status buildStatus.logEntries. forEach ( ( logE ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/names Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Template names Template names are unique identifiers used to reference and create sandboxes from your templates. They serve as human-readable names that make it easy to identify and use your templates across your applications. ​ What is a template name? A name is a string identifier that you assign to a template when building it. Once a template is built with a name, you can use that name to create sandboxes from the template. JavaScript & TypeScript Python Copy Ask AI // Build a template with a name await Template. build (template, 'my-python-env' , { cpuCount: 2 , memoryMB: 2048 , }) // Create a sandbox using the name const sandbox = await Sandbox. create ( 'my-python-env' ) ​ Team-local naming Template names are scoped to your team. This means: Your template named my-app is stored as your-team-slug/my-app You can reference it simply as my-app within your team Other teams can have their own my-app template without conflict Public templates should be referenced using the full namespaced format ( team-slug/template-name ) Backwards Compatibility : Existing public templates remain accessible without the team slug prefix. New public templates should be referenced using the full namespaced format ( team-slug/template-name ). ​ Common use cases ​ Development and production environments Use different names for different environments: JavaScript & TypeScript Python Copy Ask AI // Development template await Template. build (template, 'myapp-dev' , { cpuCount: 1 , memoryMB: 1024 , }) // Production template await Template. build (template, 'myapp-prod' , { cpuCount: 4 , memoryMB: 4096 , }) ​ Multiple template variants Create different variants of the same template with different configurations: JavaScript & TypeScript Python Copy Ask AI // Small instance await Template. build (template, 'myapp-small' , { cpuCount: 1 , memoryMB: 512 , }) // Large instance await Template. build (template, 'myapp-large' , { cpuCount: 8 , memoryMB: 8192 , }) When building variants with the same template definition but different CPU/RAM configurations, ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/tags Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Template tags & versioning Template versioning allows you to maintain multiple versions of the same template using tags. This enables workflows like semantic versioning, environment-based deployments, and gradual rollouts. ​ Tag format Tags follow the name:tag format, where name is your template’s identifier and tag is the version label. Copy Ask AI my-template:v1.0.0 // Within your team my-template:production // Within your team acme/my-template:v1.0.0 // Full namespaced reference ​ The default tag When you build or reference a template without specifying a tag, E2B uses the default tag automatically. This means: my-template is equivalent to my-template:default Existing templates without tags continue to work seamlessly JavaScript & TypeScript Python Copy Ask AI // These are equivalent const sandbox1 = await Sandbox. create ( 'my-template' ) const sandbox2 = await Sandbox. create ( 'my-template:default' ) ​ Building with tags You can build templates with one or more tags to create versioned builds. ​ Single tag JavaScript & TypeScript Python Copy Ask AI import { Template } from 'e2b' // Build with a specific version tag await Template. build (template, 'my-template:v1.0.0' ) ​ Multiple tags Build with multiple tags to assign several version labels to the same build artifact. JavaScript & TypeScript Python Copy Ask AI import { Template } from 'e2b' // Build with multiple tags pointing to the same artifact await Template. build (template, 'my-template' , { tags: [ 'v1.2.0' , 'latest' ] }) ​ Managing tags You can manage tags on existing template builds without rebuilding. ​ Assign tags Assign new tag(s) to an existing build. This is useful for promoting a tested version to production or marking a version as stable. JavaScript & TypeScript Python Copy Ask AI import { Template } from 'e2b' // Assign a single tag await Template. assignTags ( 'my-template:v1.2.0' , 'production' ) // Assign multiple tags at once await Template. assignTags ( 'my-template:v1.2.0' ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/logging Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Logging You can retrieve the build logs using the SDK. ​ Default logger We provide a default logger that you can use to filter logs by level: JavaScript & TypeScript Python Copy Ask AI import { Template, defaultBuildLogger } from 'e2b' ; await Template. build (template, 'my-template' , { onBuildLogs: defaultBuildLogger ({ minLevel: "info" , // Minimum log level to show }), }); ​ Custom logger You can customize how logs are handled: JavaScript & TypeScript Python Copy Ask AI // Simple logging onBuildLogs : ( logEntry ) => console. log (logEntry. toString ()); // Custom formatting onBuildLogs : ( logEntry ) => { const time = logEntry.timestamp. toISOString (); console. log ( `[${ time }] ${ logEntry . level . toUpperCase () }: ${ logEntry . message }` ); }; // Filter by log level onBuildLogs : ( logEntry ) => { if (logEntry.level === "error" || logEntry.level === "warn" ) { console. error (logEntry. toString ()); } }; The onBuildLogs / on_build_logs callback receives structured LogEntry objects with the following properties: JavaScript & TypeScript Python Copy Ask AI type LogEntryLevel = 'debug' | 'info' | 'warn' | 'error' class LogEntry { constructor ( public readonly timestamp : Date , public readonly level : LogEntryLevel , public readonly message : string ) toString () // Returns formatted log string } // Indicates the start of the build process class LogEntryStart extends LogEntry { constructor ( timestamp : Date , message : string ) { super (timestamp, 'debug' , message) } } // Indicates the end of the build process class LogEntryEnd extends LogEntry { constructor ( timestamp : Date , message : string ) { super (timestamp, 'debug' , message) } } Together with the LogEntry type, there are also LogEntryStart and LogEntryEnd types that indicate the start and end of the build process. Their default log level is debug and you can use them to like this: Ja ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/error-handling Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates Error handling The SDK provides specific error types: JavaScript & TypeScript Python Copy Ask AI import { AuthError, BuildError, FileUploadError } from 'e2b' ; try { await Template. build (template, 'my-template' ); } catch (error) { if (error instanceof AuthError ) { console. error ( "Authentication failed:" , error.message); } else if (error instanceof FileUploadError ) { console. error ( "File upload failed:" , error.message); } else if (error instanceof BuildError ) { console. error ( "Build failed:" , error.message); } } Was this page helpful? Yes No Suggest edits Raise issue Previous Next.js app Next.js web app running in the sandbox using Node.js Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/template/migration-v2 Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Templates V2 migration guide We’ve made ready for you three ways how to migrate existing template to the new definition format. Using the migration command (recommended) Using the fromDockerfile() method to parse existing Dockerfile Building the Docker image manually and using fromImage() method This is handled for you automatically when using any of the options below. The default user for the template is user with the working directory set to the user home directory. This is a difference from the Docker defaults. ​ Migration command (recommended) To migrate the existing template definition to the new format, follow these steps: 1 Install E2B CLI Install the latest version of the E2B CLI 2 Navigate to your template folder Navigate to the folder of your existing template definition (where you have e2b.toml and e2b.Dockerfile files). 3 Run migration command Your existing e2b.toml and e2b.Dockerfile files will be renamed to e2b.toml.old and e2b.Dockerfile.old , respectively, as they are no longer required. Copy Ask AI e2b template migrate 4 Follow the prompts Follow the prompts to complete the migration process. 5 Done ​ Generated output The migration command generates three files (based on the selected language). JavaScript & TypeScript Python Copy Ask AI template.ts - Template definition using the SDK build.dev.ts - Development build script build.prod.ts - Production build script ​ Using fromDockerfile() Method If you want to keep using Dockerfile for your template, you can use the fromDockerfile() method. We’ll automatically parse the Dockerfile for you and build the template based on it. You can find more at Base Image . JavaScript & TypeScript Python Copy Ask AI // template.ts import { Template } from 'e2b' ; const template = Template () . fromDockerfile (dockerfileContent); Only a limited set of instructions are supported and converted to equivalent template. Compatible instructions: FROM , RUN , COPY , ADD , WORKDIR , USER , ENV , ARG , CMD , ENTRYPOINT After the template definition, you can create the build scripts as described in the Quickstart section. ​ Using fromImage() Method If you already have a Docker image built, or yo ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Filesystem Each E2B Sandbox has its own isolated filesystem. The Hobby tier sandboxes come with 10 GB of the free disk space and Pro tier sandboxes come with 20 GB. With E2B SDK you can: Read and write files to the sandbox. Watch directory for changes. Upload data to the sandbox. Download data from the sandbox. Was this page helpful? Yes No Suggest edits Raise issue Previous Read & write Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem/read-write Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Read & write files ​ Reading files You can read files from the sandbox filesystem using the files.read() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const fileContent = await sandbox.files. read ( '/path/to/file' ) ​ Writing single files You can write single files to the sandbox filesystem using the files.write() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () await sandbox.files. write ( '/path/to/file' , 'file content' ) ​ Writing multiple files You can also write multiple files to the sandbox. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () await sandbox.files. write ([ { path: '/path/to/a' , data: 'file content' }, { path: '/another/path/to/b' , data: 'file content' } ]) Was this page helpful? Yes No Suggest edits Raise issue Previous File & directory metadata Next ⌘ I On this page Reading files Writing single files Writing multiple files Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem/info Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Get information about a file or directory You can get information about a file or directory using the files.getInfo() / files.get_info() methods. Information such as file name, type, and path is returned. ​ Getting information about a file JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Create a new file await sandbox.files. write ( 'test_file.txt' , 'Hello, world!' ) // Get information about the file const info = await sandbox.files. getInfo ( 'test_file.txt' ) console. log (info) // { // name: 'test_file.txt', // type: 'file', // path: '/home/user/test_file.txt', // size: 13, // mode: 0o644, // permissions: '-rw-r--r--', // owner: 'user', // group: 'user', // modifiedTime: '2025-05-26T12:00:00.000Z', // symlinkTarget: null // } ​ Getting information about a directory JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Create a new directory await sandbox.files. makeDir ( 'test_dir' ) // Get information about the directory const info = await sandbox.files. getInfo ( 'test_dir' ) console. log (info) // { // name: 'test_dir', // type: 'dir', // path: '/home/user/test_dir', // size: 0, // mode: 0o755, // permissions: 'drwxr-xr-x', // owner: 'user', // group: 'user', // modifiedTime: '2025-05-26T12:00:00.000Z', // symlinkTarget: null // } Was this page helpful? Yes No Suggest edits Raise issue Previous Watch directory for changes Next ⌘ I On this page Getting information about a file Getting information about a directory Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem/watch Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Watch sandbox directory for changes You can watch a directory for changes using the files.watchDir() method in JavaScript and files.watch_dir() method in Python. Since events are tracked asynchronously, their delivery may be delayed. It’s recommended not to collect or close watcher immediately after making a change. JavaScript & TypeScript Python Copy Ask AI import { Sandbox, FilesystemEventType } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const dirname = '/home/user' // Start watching directory for changes const handle = await sandbox.files. watchDir (dirname, async ( event ) => { console. log (event) if (event.type === FilesystemEventType. WRITE ) { console. log ( `wrote to file ${ event . name }` ) } }) // Trigger file write event await sandbox.files. write ( `${ dirname }/my-file` , 'hello' ) ​ Recursive watching You can enable recursive watching using the parameter recursive . When rapidly creating new folders (e.g., deeply nested path of folders), events other than CREATE might not be emitted. To avoid this behavior, create the required folder structure in advance. JavaScript & TypeScript Python Copy Ask AI import { Sandbox, FilesystemEventType } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const dirname = '/home/user' // Start watching directory for changes const handle = await sandbox.files. watchDir (dirname, async ( event ) => { console. log (event) if (event.type === FilesystemEventType. WRITE ) { console. log ( `wrote to file ${ event . name }` ) } }, { recursive: true }) // Trigger file write event await sandbox.files. write ( `${ dirname }/my-folder/my-file` , 'hello' ) Was this page helpful? Yes No Suggest edits Raise issue Previous Upload data Next ⌘ I On this page Recursive watching Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem/upload Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Upload data to sandbox You can upload data to the sandbox using the files.write() method. ​ Upload single file JavaScript & TypeScript Python Copy Ask AI import fs from 'fs' import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Read file from local filesystem const content = fs. readFileSync ( '/local/path' ) // Upload file to sandbox await sandbox.files. write ( '/path/in/sandbox' , content) ​ Upload with pre-signed URL Sometimes, you may want to let users from unauthorized environments, like a browser, upload files to the sandbox. For this use case, you can use pre-signed URLs to let users upload files securely. All you need to do is create a sandbox with the secure: true option. An upload URL will then be generated with a signature that allows only authorized users to upload files. You can optionally set an expiration time for the URL so that it will be valid only for a limited time. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' // Start a secured sandbox (all operations must be authorized by default) const sandbox = await Sandbox. create (template, { secure: true }) // Create a pre-signed URL for file upload with a 10 second expiration const publicUploadUrl = await sandbox. uploadUrl ( 'demo.txt' , { useSignatureExpiration: 10_000 , // optional }, ) // Upload a file with a pre-signed URL (this can be used in any environment, such as a browser) const form = new FormData () form. append ( 'file' , 'file content' ) await fetch (publicUploadUrl, { method: 'POST' , body: form }) // File is now available in the sandbox and you can read it const content = sandbox.files. read ( '/path/in/sandbox' ) ​ Upload directory / multiple files JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Read all files in the directory and store their paths and contents in an array const readDirectoryFiles = ( ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/filesystem/download Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Filesystem Download data from sandbox You can download data from the sandbox using the files.read() method. JavaScript & TypeScript Python Copy Ask AI import fs from 'fs' import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Read file from sandbox const content = await sandbox.files. read ( '/path/in/sandbox' ) // Write file to local filesystem fs. writeFileSync ( '/local/path' , content) ​ Download with pre-signed URL Sometimes, you may want to let users from unauthorized environments, like a browser, download files from the sandbox. For this use case, you can use pre-signed URLs to let users download files securely. All you need to do is create a sandbox with the secure: true option. A download URL will then be generated with a signature that allows only authorized users to access files. You can optionally set an expiration time for the URL so that it will be valid only for a limited time. JavaScript & TypeScript Python Copy Ask AI import fs from 'fs' import { Sandbox } from '@e2b/code-interpreter' // Start a secured sandbox (all operations must be authorized by default) const sandbox = await Sandbox. create (template, { secure: true }) // Create a pre-signed URL for file download with a 10 second expiration const publicUrl = await sandbox. downloadUrl ( 'demo.txt' , { useSignatureExpiration: 10_000 , // optional }, ) // Download a file with a pre-signed URL (this can be used in any environment, such as a browser) const res = await fetch (publicUrl) const content = await res. text () Was this page helpful? Yes No Suggest edits Raise issue Previous Overview Next ⌘ I On this page Download with pre-signed URL Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/commands Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Commands Running commands in sandbox You can run terminal commands inside the sandbox using the commands.run() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const result = await sandbox.commands. run ( 'ls -l' ) console. log (result) Was this page helpful? Yes No Suggest edits Raise issue Previous Streaming Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/commands/streaming Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Commands Streaming command output To stream command output as it is being executed, pass the onStdout , onStderr callbacks to the commands.run() method in JavaScript or the on_stdout , on_stderr callbacks to the commands.run() method in Python. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () const result = await sandbox.commands. run ( 'echo hello; sleep 1; echo world' , { onStdout : ( data ) => { console. log (data) }, onStderr : ( data ) => { console. log (data) }, }) console. log (result) Was this page helpful? Yes No Suggest edits Raise issue Previous Run commands in background Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/commands/background Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation Commands Running commands in background To run commands in background, pass the background option to the commands.run() method. This will return immediately and the command will continue to run in the sandbox. You can then later kill the command using the commands.kill() method. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from '@e2b/code-interpreter' const sandbox = await Sandbox. create () // Start the command in the background const command = await sandbox.commands. run ( 'echo hello; sleep 10; echo world' , { background: true , onStdout : ( data ) => { console. log (data) }, }) // Kill the command await command. kill () Was this page helpful? Yes No Suggest edits Raise issue Previous Overview Connect to 200+ tools through the Model Context Protocol Next ⌘ I Assistant Responses are generated using AI and may contain mistakes. --- ## Documentation - E2B > Source: https://e2b.dev/docs/mcp Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation MCP gateway Overview E2B provides a batteries-included MCP gateway that runs inside sandboxes, giving you type-safe access to 200+ MCP tools from the Docker MCP Catalog or custom MCPs through a unified interface. This integration gives developers instant access to tools like Browserbase , Exa , Notion , Stripe , or GitHub . The Model Context Protocol (MCP) is an open standard for connecting AI models to external tools and data sources. E2B sandboxes provide an ideal environment for running MCP tools, giving AI full access to an internet-connected Linux machine where it can safely install packages, write files, run terminal commands, and AI-generated code. TypeScript Python Copy Ask AI import Sandbox from 'e2b' const sbx = await Sandbox. create ({ mcp: { browserbase: { apiKey: process.env. BROWSERBASE_API_KEY ! , geminiApiKey: process.env. GEMINI_API_KEY ! , projectId: process.env. BROWSERBASE_PROJECT_ID ! , }, exa: { apiKey: process.env. EXA_API_KEY ! , }, airtable: { airtableApiKey: process.env. AIRTABLE_API_KEY ! , }, }, }); const mcpUrl = sbx. getMcpUrl (); const mcpToken = await sbx. getMcpToken (); // You can now connect the gateway to any MCP client, for example claude: // This also works for your local claude! await sbx.commands. run ( `claude mcp add --transport http e2b-mcp-gateway ${ mcpUrl } --header "Authorization: Bearer ${ mcpToken }"` , { timeoutMs: 0 , onStdout: console.log, onStderr: console.log }); await sbx.commands. run ( `echo 'Use browserbase and exa to research open positions at e2b.dev. Collect your findings in Airtable.' | claude -p --dangerously-skip-permissions` , { timeoutMs: 0 , onStdout: console.log, onStderr: console.log } ) ​ Documentation Quickstart Get started with MCP Available servers Browse 200+ pre-built MCP servers Custom templates Prepull MCP servers for faster runtime Custom servers Use custom MCP servers from GitHub Examples See examples Was this page helpful? Yes No Suggest edits Raise issue Previous Quickstart Get started with MCP integra ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/mcp/quickstart Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation MCP gateway Quickstart You can connect to the MCPs running inside the sandbox both from outside and inside the sandbox. ​ From outside the sandbox To connect to the MCPs running inside the sandbox, use the sandbox.getMcpUrl() in JavaScript and sandbox.get_mcp_url() in Python. TypeScript Python Copy Ask AI import Sandbox from 'e2b' import { Client } from '@modelcontextprotocol/sdk/client/index.js' ; import { StreamableHTTPClientTransport } from '@modelcontextprotocol/sdk/client/streamableHttp.js' ; const sandbox = await Sandbox. create ({ mcp: { browserbase: { apiKey: process.env. BROWSERBASE_API_KEY ! , geminiApiKey: process.env. GEMINI_API_KEY ! , projectId: process.env. BROWSERBASE_PROJECT_ID ! , }, exa: { apiKey: process.env. EXA_API_KEY ! , }, notion: { internalIntegrationToken: process.env. NOTION_API_KEY ! , }, }, }); const client = new Client ({ name: 'e2b-mcp-client' , version: '1.0.0' }); const transport = new StreamableHTTPClientTransport ( new URL (sandbox. getMcpUrl ()), { requestInit: { headers: { 'Authorization' : `Bearer ${ await sandbox . getMcpToken () }` } } } ); await client. connect (transport); const tools = await client. listTools (); console. log ( 'Available tools:' , tools.tools. map ( t => t.name)); await client. close (); await sandbox. kill (); ​ From inside the sandbox If you need to access the MCP gateway from within the sandbox itself, it’s available at: Copy Ask AI http://localhost:50005/mcp You’ll need to include the Authorization header with the MCP token when making requests from inside the sandbox. How that is added depends on the MCP client you use: ​ Claude Copy Ask AI claude mcp add --transport http e2b-mcp-gateway --header "Authorization: Bearer " ​ OpenAI Agents TypeScript Python Copy Ask AI import { MCPServerStreamableHttp } from  ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/mcp/available-servers Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation MCP gateway Available servers E2B provides access to 200+ MCP servers from Docker’s catalog . You can also run custom MCP servers inside the sandbox. ​ Airtable Provides AI assistants with direct access to Airtable bases, allowing them to read schemas, query records, and interact with your Airtable data. Supports listing bases, retrieving table structures, and searching through records to help automate workflows and answer questions about your organized data. View on Docker Hub ​ airtable object Show properties ​ airtableApiKey string required ​ nodeenv string required ​ Azure Kubernetes Service (AKS) Azure Kubernetes Service (AKS) official MCP server. View on Docker Hub ​ aks object Show properties ​ accessLevel string required Access level for the MCP server, One of [ readonly, readwrite, admin ] ​ additionalTools string Comma-separated list of additional tools, One of [ helm, cilium ] ​ allowNamespaces string Comma-separated list of namespaces to allow access to. If not specified, all namespaces are allowed. ​ azureDir string required Path to the Azure configuration directory (e.g. /home/azureuser/.azure). Used for Azure CLI authentication, you should be logged in (e.g. run az login ) on the host before starting the MCP server. ​ containerUser string Username or UID of the container user (format [: ] e.g. 10000), ensuring correct permissions to access the Azure and kubeconfig files. Leave empty to use default user in the container. ​ kubeconfig string required Path to the kubeconfig file for the AKS cluster (e.g. /home/azureuser/.kube/config). Used to connect to the AKS cluster. ​ Apify Apify is the world’s largest marketplace of tools for web scraping, data extraction, and web automation. You can extract structured data from social media, e-commerce, search engines, maps, travel sites, or any other website. View on Docker Hub ​ apify object Show properties ​ apifyToken string required ​ tools string required Comma-separated list of tools to enable. Can be either a tool category, a specific tool, or an Apify Actor. For example: “actors,docs,apify/rag-web-browser”. For more details visit https://mcp.apify.com . ​ Api-gateway A universal MCP (Mod ...(truncated) --- ## Documentation - E2B > Source: https://e2b.dev/docs/mcp/custom-templates Documentation - E2B Skip to main content E2B Docs home page Search... ⌘ K Documentation SDK reference Getting started Home Quickstart API key Cookbook Need help? Billing Use cases Coding Agents Computer use GitHub Actions CI/CD Agents in sandbox Amp Claude Code Codex OpenClaw OpenCode Code interpreting Analyze data with AI Charts & visualizations Streaming Code contexts Supported languages Sandbox Lifecycle Lifecycle events API Lifecycle events webhooks Persistence Snapshots Git integration Metrics Metadata Environment variables List sandboxes Connect to running sandbox Internet access Interactive terminal SSH access Connecting storage bucket Rate limits Secured access Proxy tunneling Custom domain Templates Quickstart How it works User and workdir Caching Base image Private registries Defining template Start & ready commands Build Names Tags & versioning Logging Error handling Examples V2 migration guide Filesystem Overview Read & write File & directory metadata Watch directory for changes Upload data Download data Commands Overview Streaming Run commands in background MCP gateway Overview Quickstart Available servers Custom templates Custom servers Examples CLI Installation Authentication List sandboxes Create sandbox Connect to sandbox Execute commands in sandbox Shutdown running sandboxes Deployment Bring Your Own Cloud Migration SDK v2 migration guide Troubleshooting Build authentication error The sandbox is running but port is not open - 49999 pip install fails during template build Dashboard e2b-dev/e2b E2B Docs home page Search... ⌘ K Ask AI Search... Navigation MCP gateway Custom templates You can prepull MCP server Docker images during template build time to significantly improve runtime performance. ​ How it works When you build a template with prepulled MCP servers, the Docker images for those servers are downloaded and cached during the build process. This means when you create a sandbox from that template, the MCP servers are ready to use immediately without waiting for image downloads. You must use the MCP gateway enabled template ( mcp-gateway ) as your base template to use this feature. ​ Building a template with MCP servers Use the addMcpServer() method (TypeScript) or add_mcp_server() method (Python) to prepull MCP server images during template build. You can pass a single server or an array of servers. The server names (like "browserbase" and "exa" ) correspond to the keys defined in the Available Servers documentation. JavaScript & TypeScript Python Copy Ask AI import "dotenv/config" ; import { Template, defaultBuildLogger } from 'e2b' ; export const template = Template () . fromTemplate ( "mcp-gateway" ) . addMcpServer ([ "browserbase" , "exa" ]); await Template. build (template, 'my-mcp-gateway' , { cpuCount: 8 , memoryMB: 8192 , onBuildLogs: defaultBuildLogger (), }); ​ Using the template Once built, create sandboxes from your template. You still need to provide the configuration for each MCP server. JavaScript & TypeScript Python Copy Ask AI import { Sandbox } from 'e2b' ; const sandbox = await Sandbox. create ({ template: "my-mcp-gateway" , mcp: { browserbase: { apiKey: process.env. BROWSERBASE_API_KEY ! , geminiApiKey: process.env. GEMINI_API_KEY ! , projectId: process.env. BROWSERBASE_PROJECT_ID ! , }, exa: { apiKey: process.env. EXA_API_KEY ! , }, }, }); ​ Learn more For more information about working with templates, see: Template Quickstart - Get started with custom templates Defining Templates - Learn all template configuration options Template Build - Understand the build process Was this page helpful? Yes No Sugges ...(truncated) ---