Sandbox Lifecycle

Creating a Sandbox

Python
TypeScript

from synapse import Cell

# Basic — ephemeral sandbox
cell = Cell()

# With options
cell = Cell(
    template="python3",
    persistent=True,
    timeout_ms=3_600_000,  # 1 hour
    metadata={"project": "my-agent"},
    envs={"API_KEY": "sk_..."},
)

# Factory method (async context)
cell = Cell.create(template="python3")

import { Cell } from '@runsynapse/sdk';

const cell = await Cell.create({
  template: 'python3',
  persistent: true,
  timeoutMs: 3_600_000,
  metadata: { project: 'my-agent' },
  envs: { API_KEY: 'sk_...' },
});

Connecting to an Existing Sandbox

Python
TypeScript

cell = Cell.connect("cell_id_here")

const cell = await Cell.connect("cell_id_here");

Listing Sandboxes

Python
TypeScript

cells = Cell.list()
for info in cells:
    print(f"{info.sandbox_id}: {info.state}")

const paginator = Cell.list({ limit: 50 });
while (paginator.hasNext) {
  const items = await paginator.nextItems();
  items.forEach(info => console.log(`${info.sandboxId}: ${info.state}`));
}

Pause and Resume

Sandbox state is preserved across pause/resume cycles:

# Pause — creates a filesystem snapshot
snapshot = cell.pause()
print(snapshot.snapshot_id)

# Resume — restores the snapshot
cell.resume()

Timeout Management

# Set timeout (seconds)
cell.set_timeout(7200)  # 2 hours

# Reset inactivity timer
cell.keep_alive(3600)

Kill

cell.kill()  # Destroys the sandbox and frees resources