E2B MCP
Give any AI agent a secure cloud sandbox to run code in
Last verified: August 31, 2026
How to Install E2B MCP
$npx -y @e2b/mcp-serverRequires Claude Desktop, Cursor, Windsurf, or another MCP-compatible client.
About E2B MCP
E2B's MCP server exposes Code Interpreter capabilities to any MCP client, giving an agent an isolated cloud sandbox where it can execute Python and JavaScript, run shell commands, and drive a headless browser. Because the sandbox is fully separated from your machine and your production systems, agents can run untrusted or generated code without putting local files at risk. It is the standard choice when you want an agent that writes code to also be able to run and verify it.
E2B MCP is a Code Execution MCP server designed for data scientists, engineers, and power users. It enables giving AI assistants a sandboxed runtime to actually run code by providing isolated Python/Node/Bash execution as MCP tools. Key capabilities include sandboxed python, node, or bash, file upload and download, cpu and memory caps, and structured tool calls that work cleanly inside an LLM context. It integrates with Python, Node.js, Bash, Deno, and common data libraries, and is best suited for data scientists, engineers, and power users who need data analysis on ad-hoc files.
Key Features
- code-execution
- sandbox
- python
- javascript
- security
Pricing
- Core MCP server
- Community support
- Works with any MCP client
- Everything in Free
- Higher usage limits
- Priority support
- Everything in Pro
- SLA & SSO
- Dedicated support
Tier details are indicative — visit the E2B MCP website for current pricing.
Pros & Cons
Pros
- Verifies output instead of guessing
- Unlocks data analysis on uploaded files
- Strong isolation keeps the host safe
Cons
- Latency overhead vs in-process execution
- Network restrictions can limit some workflows
- Stateless model means re-uploading context
Best For
- Data analysis on ad-hoc files
- Repro from a stack trace
- Script generation with verification
- One-shot ETL jobs
Screenshots
Screenshots coming soon —
Submit yours →Our Take on E2B MCP
The default pick when your agent needs to actually run the code it writes. The isolation story is the reason to use it — running generated code locally is how you lose a weekend.