Skip to main content
You are reading a legacy (pre v1.0) document.
Custom sandboxes allow you to spawn a sandbox with an environment predefined by you and then control it with our SDK.
The E2B Sandbox is a secure way to run your AI app. It is a long-running cloud environment where you can let any LLM (GPTs, Claude, local LLMs, etc) use tools exactly like you would do locally.Read more about E2B Sandboxes here.

​
How Custom Sandboxes Work

Follow our guide on how to create a custom sandbox.
  1. You build a custom sandbox by creating the Sandbox Template which is a Dockerfile describing the environment of your sandbox.
  2. Build a sandbox template from it using our CLI
  3. Get a template ID from the CLI: 
    # ... truncated CLI output
✅ Building sandbox template 3y5bvra6kgq0kaumgztu finished.
  4. Pass the template ID to our SDK like this: 
    import { Sandbox } from 'e2b'

// Create new sandbox
const sandbox = await Sandbox.create({
  // You get sandbox template from the CLI after you run `$ e2b template build`
  template: '<sandbox-template>',
})

// Close sandbox once done
await sandbox.close()
  5. Run the code and we spawn a sandbox based on the template you built in step 2
Once you build your custom sandbox template, you can spawn multiple isolated sandboxes from it. You can repeat the steps 4, 5 multiple times using the same ID, always creating a separate sandbox instance.
Follow our guide on how to create a custom sandbox.