Sandboxes

Running code

Agents

Learn E2B sandboxes from the SDK's source: run_code and Execution, errors, results, files, network security, and a code tool for agents tested without a key.

E2B overview

E2B setup: API key and first sandbox

Installation and setup

E2B sandbox lifecycle and timeouts

Sandbox lifecycle: templates, timeouts and kill

E2B run_code and Execution results

run_code and the Execution object

E2B code errors and timeouts

Errors: reported, not raised

E2B results, charts and display formats

Results, displays and charts

E2B filesystem and commands

Files and shell commands

E2B sandbox security: internet access and secrets

Security settings: network, secrets and timeouts

E2B code interpreter tool for LLM agents

A code tool for an LLM, tested without a sandbox

E2B data analysis agent with a code tool

Data analysis agent with a tested tool

Agents quiz

What should a sandbox for model-written code usually not have?