Runpod, a cloud company for AI developers, released an open-source Python SDK called Flash to reduce the infrastructure burden in deploying AI code to production environments, Techzine reported on Sunday.

The company said Flash helps developers turn local Python functions into autoscaling endpoints in minutes without building containers, managing images or configuring infrastructure. It is available under the MIT license on PyPI and GitHub.

Runpod CEO Zhen Lu (젠 루) said, "Serverless is powerful, but we have consistently received feedback that the setup process is a stumbling block." He said, "The goal is for developers to write Python code and choose compute, and be able to handle requests within minutes."

He added, "Flash is well-suited to these workloads, as agents need to call different models, move between different compute types, and scale based on demand."

Flash supports two deployment methods: queue-based processing for batch and asynchronous workloads, and load-balancing endpoints for real-time inference traffic.

When developers specify compute requirements and dependencies directly in Python, Flash automatically handles provisioning, scaling and infrastructure management. Endpoints automatically scale based on demand and scale down to zero when there is no usage. The company said more than 700,000 developers currently use Runpod, and 37,000 serverless endpoints were created in March 2026 alone.