Use this file to discover all available pages before exploring further.
The sandbox persistence allows you to pause your sandbox and resume it later from the same state it was in when you paused it.This includes not only state of the sandbox’s filesystem but also the sandbox’s memory. This means all running processes, loaded variables, data, etc.
Please note:
It takes about 4 seconds per 1 GB RAM to pause the sandbox.
It takes about 1 second to resume the sandbox.
The data of a paused sandbox is permanently retained until you explicitly call the kill method.
When you pause a sandbox, both the sandbox’s filesystem and memory state will be saved. This includes all the files in the sandbox’s filesystem and all the running processes, loaded variables, data, etc.
import { Sandbox } from 'novita-sandbox/code-interpreter'const sandbox = await Sandbox.create()console.log('Sandbox created', sandbox.sandboxId)// Pause the sandbox.// You can save the sandbox ID in your database// to resume the sandbox laterconst result = await sandbox.betaPause()console.log('Sandbox paused', sandbox.sandboxId, result)await sandbox.kill()
Note: Brief transition period after pausingPausing is an asynchronous operation. After calling beta_pause(), the sandbox pauses in the background.During this transition period, before the pause completes:
Read-only operations (e.g., get_info, list): return normally; status is shown as paused
Execution operations (e.g., commands.run, accessing sandbox public URLs): will return errors — this is expected behavior. Wait for pausing to complete before retrying (time varies with memory size, see note above)
Resume operation (connect): automatically waits for the pause to complete before resuming the sandbox — no manual polling or retry needed
When you resume a sandbox, it will be in the same state it was in when you paused it.
This means that all the files in the sandbox’s filesystem will be restored and all the running processes, loaded variables, data, etc. will be restored.
If you call connect() immediately after calling beta_pause(), connect() automatically waits for the pause to complete before resuming the sandbox. No manual polling is required — the entire process is transparent to the caller.
If you try to resume a sandbox that has been destroyed or does not exist:
JavaScript SDK will throw NotFoundError
Python SDK will throw NotFoundException
import { Sandbox } from 'novita-sandbox/code-interpreter'const sandbox = await Sandbox.create()console.log('Sandbox created', sandbox.sandboxId)// Pause the sandbox.// You can save the sandbox ID in your database// to resume the sandbox laterconst result = await sandbox.betaPause()console.log('Sandbox paused', sandbox.sandboxId, result)// Resume the sandbox.// Even if called before pausing completes, connect() automatically waits for the pause to finish.const resumedSandbox = await sandbox.connect()console.log('Sandbox resumed', resumedSandbox.sandboxId)await sandbox.kill()
You can list all paused sandboxes by calling the Sandbox.list method and supplying the state query parameter.
More information about using the method can be found in List Sandboxes.
import { Sandbox, SandboxInfo } from 'novita-sandbox/code-interpreter'const sandbox = await Sandbox.create()// List all paused sandboxes.const paginator = Sandbox.list({ query: { state: ['paused'] } })// Get all paused sandboxes.const sandboxes: SandboxInfo[] = []while (paginator.hasNext) { const items = await paginator.nextItems() sandboxes.push(...items)}console.log('all paused sandboxes', sandboxes)await sandbox.kill()
When you resume a sandbox, the sandbox’s timeout is reset to the default timeout of a sandbox - 5 minutes.You can pass a custom timeout to the Sandbox.connect() method like this:
If you have a service (for example a server) running inside your sandbox and you pause the sandbox, the service won’t be accessible from the outside and all the clients will be disconnected.
If you resume the sandbox, the service will be accessible again but you need to connect clients again.