Favicon of E2B

E2B

AI agent sandbox infrastructure with isolated Linux machines, an Apache 2.0 runtime and hosted or self-hosted deployment.

Screenshot of E2B website

E2B provides isolated execution environments for developers building coding assistants, computer-use agents and background jobs. You supply the model and agent logic; it supplies the machine where the agent works.

  • Isolated Linux machines give each agent session its own Firecracker-based microVM with a dedicated kernel, shell, filesystem and browser.
  • Code execution runs Python or JavaScript through Code Interpreter and returns outputs such as charts, tables and images.
  • Desktop control gives agents a screen, mouse and keyboard, with a live stream so you can observe their actions.
  • Saved environments support pausing, resuming, snapshots and forks, alongside persistent files.
  • Network controls govern access to model providers, MCP tools, repositories, storage buckets and internal APIs.

The open-source runtime uses Apache 2.0. E2B Embed lets you self-host the full stack on a Linux machine, with Docker Compose or Kubernetes deployment options. E2B Cloud is a paid managed service that requires an account and API key. E2B also operates deployments in your AWS, GCP or Azure account, where sandboxes and data stay in your VPC.

Python and JavaScript SDKs connect the environments to agent applications. Supported stacks include OpenAI Agents SDK, Claude Code and Codex. Agents can use Anthropic, OpenAI or open-weight models; their external requests follow the sandbox's network policy.

Similar to E2B