e2b-mcp-server
Production-grade MCP server that enables AI assistants to execute code securely in isolated E2B sandboxes.
README

E2B MCP Server
Production-grade Model Context Protocol servers that enable AI assistants to execute code securely in isolated E2B sandboxes. Bring powerful code interpretation capabilities to Claude Desktop, Cline, and other MCP-compatible clients.
Demo • JavaScript Docs • Python Docs • E2B Platform
Overview
This repository contains production-ready MCP server implementations in both JavaScript/TypeScript and Python. Each server provides a complete toolkit for AI models to interact with secure, isolated code execution environments.
What you can do:
- Execute Python code in Jupyter-style notebooks
- Run shell commands in isolated environments
- Manage files (read, write, list) within sandboxes
- Expose services running in sandboxes via public URLs
- Manage multiple concurrent sandboxes with automatic cleanup
- Monitor sandbox usage with comprehensive logging
Key features:
- 10 Production Tools: Complete sandbox interaction toolkit
- Resource Management: Automatic lifecycle management with configurable limits
- Type Safety: Full validation with Zod (JS) and Pydantic (Python)
- Robust Error Handling: Clear diagnostics and graceful degradation
- Production Logging: Multi-level logging with timestamps
- Graceful Shutdown: Proper cleanup on termination signals
Quick Start
Manual Installation
JavaScript/TypeScript
# With npx (no installation required)
npx @yukkit/e2b-mcp-server
# Or install globally
npm install -g @yukkit/e2b-mcp-server
Python
# With uv (recommended)
uv pip install e2b-mcp-server
# With pip
pip install e2b-mcp-server
Configuration
Claude Desktop
Add to your Claude Desktop config:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"e2b": {
"command": "npx",
"args": ["-y", "@yukkit/e2b-mcp-server"],
"env": {
"E2B_API_KEY": "your-api-key-here"
}
}
}
}
Environment Variables
| Variable | Description | Default |
|---|---|---|
E2B_API_KEY |
Your E2B API key (get one here) | Required |
MAX_ACTIVE_SANDBOXES |
Maximum concurrent sandboxes | 10 |
LOG_LEVEL |
Logging level (DEBUG, INFO, WARNING, ERROR) | INFO |
[!IMPORTANT] Get your free E2B API key at e2b.dev
Available Tools
| Tool | Description | Parameters |
|---|---|---|
| create_sandbox | Create a new isolated sandbox | secure (default: true), timeoutMs (optional) |
| run_code | Execute Python code in a sandbox | code, sandboxId (optional) |
| run_command | Run shell commands | command, sandboxId, background |
| read_file | Read file contents | filePath, sandboxId |
| write_file | Write to a file | filePath, fileContents, sandboxId |
| list_files | List directory contents | folderPath, sandboxId |
| get_sandbox_url | Get public URL for a port | port, sandboxId |
| get_file_download_url | Get download URL for a file | filePath, sandboxId, useSignatureExpiration (default: 300000ms / 5 min) |
| list_sandbox_ids | List all active sandboxes | None |
| kill_sandbox | Terminate a sandbox | sandboxId |
Language Support
Choose the implementation that fits your stack:
JavaScript/TypeScript
- Built with TypeScript for type safety
- Uses Zod for input validation
- Compatible with Node.js 20+
- Full documentation: packages/js/README.md
Install:
npm install @yukkit/e2b-mcp-server
Python
- Type-safe with full type annotations
- Pydantic models for validation
- Python 3.10+ required
- Full documentation: packages/python/README.md
Install:
pip install e2b-mcp-server
Development
This is a monorepo managed with pnpm workspaces.
Prerequisites
- Node.js 20+
- pnpm 9+
- Python 3.10+ (for Python package)
- Poetry (for Python package)
Setup
# Clone the repository
git clone https://github.com/e2b-dev/mcp-server.git
cd mcp-server
# Install dependencies
pnpm install
# Build all packages
pnpm build
Project Structure
.
├── packages/
│ ├── js/ # TypeScript/JavaScript implementation
│ │ ├── src/ # Source code
│ │ └── build/ # Compiled output
│ └── python/ # Python implementation
│ └── e2b_mcp_server/ # Package source
├── .changeset/ # Changesets for version management
└── .github/ # CI/CD workflows
Publishing
This project uses Changesets for version management:
# Create a changeset
pnpm changeset
# Version packages
pnpm run version
# Publish to npm/PyPI
pnpm run publish
Use Cases
AI-Powered Code Assistants: Let AI models write and execute code with immediate feedback
Data Analysis: Enable AI to analyze datasets, create visualizations, and generate reports
DevOps Automation: Allow AI to interact with systems and run diagnostic commands
Educational Tools: Create interactive coding tutorials with safe execution environments
API Testing: Let AI test and validate APIs by running actual requests
Code Generation & Testing: Generate code and verify it works through execution
Resources
- E2B Documentation
- Model Context Protocol
- E2B Code Interpreter
- MCP TypeScript SDK
- Claude Desktop
- Smithery MCP Registry
Support
- Documentation: Check the JavaScript or Python READMEs
- Issues: Open an issue
- E2B Platform: support@e2b.dev
- Discord: Join the E2B Community
Security
E2B sandboxes provide secure, isolated execution environments. However, always:
- Keep your E2B API key secure and never commit it to version control
- Use environment variables for sensitive configuration
- Monitor sandbox usage and set appropriate limits
- Review code before execution in production environments
For security concerns, please email security@e2b.dev.
Contributing
Contributions are welcome! Please read our contributing guidelines before submitting PRs.
- Fork the repository
- Create a feature branch
- Make your changes with tests
- Create a changeset:
pnpm changeset - Submit a pull request
License
Apache 2.0 - See LICENSE for details.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。