Keep your agent where it is.
Let it use a Hooklet computer to run commands and tools.
Cloud computers for AI agents.
Run code. Automate work.
Hooklet gives your agents isolated cloud computers to run code, use tools, and work in parallel.
Developers building on Hooklet work at these companies
A workspace for every agent.
Give each task its own computer to run commands, install packages, and work with files.
Connect your agent. Or host it here.
Run it on Hooklet.
Put the agent and its tools in the same cloud workspace.
Run tasks in parallel.
Split a code review across four workspaces. Check security, tests, performance, and dependencies at the same time.
Start work from a webhook.
Planned: trigger a task when an event arrives. A bug report comes in; an agent investigates and opens a pull request.
Keep each task isolated.
Each workspace runs in a separate microVM. Stop or restart it without interrupting the others.
Launch agent workspaces from your workflow.
Use the Python SDK to create a cloud workspace, run your agent’s code, and stop the workspace when the task is done. Build it into your app, job, or pipeline.
Explore the docsPython SDK
import asyncio
from agent_sandbox import Sandbox
async def main():
vm = await Sandbox.create("my-agent")
try:
result = await vm.exec("python", ["-c", "print('Hello')"])
print(result.stdout_text)
finally:
await vm.stop()
asyncio.run(main())Requires configured infrastructure. Setup guide
