Integrations

Run Team Pool workers on an infrastructure partner's platform, or start from a reference template you clone and adapt. Every option runs the same worker: the Cursor CLI opens an outbound HTTPS connection to Cursor, and Cursor sends agent tool calls over that connection.

Partner guides and templates are reference architectures. You own the worker image, infrastructure, secrets, scaling policy, and production validation.

Partner guides

Each partner maintains its own guide for running Self-Hosted Machines workers on its platform:

• AWS Lambda. Cursor Self-Hosted Machines on Lambda MicroVMs
• Cloudflare. Cursor Cloud Agents on Cloudflare Sandbox
• Namespace. Cursor integration
• Modal. Cursor on Modal
• Daytona. Cursor Self-Hosted Machines on Daytona
• E2B. Cursor agents on E2B
• Vercel. Cursor with Vercel Sandbox

Reference templates

Clone one of these repositories to start from a working worker deployment. Each one runs a worker controller that claims pending pool requests and starts one worker per claim:

• AWS Lambda MicroVMs. anysphere/aws-lambda-workers: a --spawn hook launches one Firecracker-isolated Lambda MicroVM per claimed request.
• Cloudflare Containers. anysphere/cloudflare-workers: a Cloudflare Worker acts as the controller and starts one Cloudflare Container per claimed request.
• Kubernetes. anysphere/k8s-workers: a --spawn sample that creates a Pod per claimed request, or keeps warm idle Pods, without a CRD. For the operator path, where a Helm chart manages WorkerDeployment resources and warm capacity, follow the Kubernetes guide.

What every integration needs

The platform changes, but the worker requirements stay the same:

• A Cursor Enterprise plan, with Self-Hosted Machines enabled by a team admin in the Cloud Agents dashboard.
• A service account API key in CURSOR_API_KEY. See Authenticate workers.
• Outbound HTTPS to api2.cursor.sh, api2direct.cursor.sh, and cloud-agent-artifacts.s3.us-east-1.amazonaws.com. See Networking.
• A pool name or labels so Cursor routes the right requests to workers on that platform.

For a single personal machine instead of a platform-managed fleet, use My Machines.

Next steps

• Team Pools: start a pool worker by hand before you automate it on a platform.
• Worker controller: the --spawn hook model every template builds on.
• Kubernetes: operator and Helm chart for pools at scale.
• API reference: endpoints for workers, pools, the pending-request queue, and worker tokens.