E2B Code Sandbox

The Jupyter MCP Server can execute code in an E2B sandbox, an isolated cloud environment running a Jupyter kernel. Execution is routed through the code-sandboxes e2b engine.

tip

For the full, engine-level credential and parameter reference, see the E2B sandbox guide in the code-sandboxes documentation.

Requirements

Install the E2B extra:

pip install "jupyter-mcp-server[e2b]"

Provide credentials through environment variables:

Environment variable	Description
`E2B_API_KEY`	E2B API key
`E2B_DOMAIN`	E2B deployment to reach, when it is not the default `e2b.dev`

The sandbox is created from E2B's code-interpreter-v1 template — the single environment this engine offers, e2b-code-interpreter. That template is the one carrying the Jupyter kernel the code interpreter talks to; one without a kernel would start and then fail every execution.

Configuration

Select the E2B engine with SANDBOX_VARIANT=e2b:

SANDBOX_VARIANT=e2b
E2B_API_KEY=your-e2b-api-key

Or via the command line:

jupyter mcp start \
  --transport streamable-http \
  --sandbox-variant e2b \
  --port 4040

MCP client configuration:

{
  "mcpServers": {
    "jupyter": {
      "command": "uvx",
      "args": ["jupyter-mcp-server@latest"],
      "env": {
        "SANDBOX_VARIANT": "e2b",
        "E2B_API_KEY": "your-e2b-api-key"
      }
    }
  }
}

note

Each execution context is a Jupyter kernel, so variables, imports and definitions persist across executions, and rich outputs — figures, HTML — come back as results, as they would from a local kernel.

Requirements​

Configuration​

Requirements

Configuration