Run Python remotely, then ship it as an endpoint.
serverless inference
Runs Python functions on remote Runpod GPUs or CPUs during development and deploys them as serverless endpoints.
When to use it
Use it for @Endpoint or @remote functions, resource configuration, and debugging Flash deployments.
Give it a Python project and endpoint requirements; it returns code and can run it with `flash dev` or deploy it with `flash deploy`.
What you provide
This skill
Runpod serverless endpoints
Deploys serverless endpoints (irreversible)
Deletes deployed endpoints (irreversible)
Runpod remote workers
Functions on remote workers
flash dev server
Starts the development server
Flash development logs
Reads live worker logs
The runpod-flash CLI must be installed; supported Python versions are 3.10 through 3.13, not 3.14+.
A Runpod account is required for authentication and endpoint provisioning.