Developers build a serverless photo gallery on Google Cloud that automatically transforms uploaded images into stylized art. When a file lands in a temporary bucket, a Cloud Run function triggers, calls the Vision API to label the picture, and stores metadata in a Firestore database. The same function then pulls the image into memory, sends it to Vertex AI’s Gemini Nano Banana model with a cyber‑punk prompt, and receives a PNG back.

The augmented image is uploaded to a permanent bucket, and its public URL is added to the Firestore record. A lightweight website, also running on Cloud Run, reads the Firestore collection and renders a gallery page with image URLs and labels. Terraform scripts provision the Firestore instance, Cloud Functions, IAM roles, and bucket permissions, keeping the deployment reproducible.

The result is a fully automated pipeline that turns ordinary photos into themed artwork and displays them instantly on a web page. Future extensions could add video generation or audio narration, but the core stack already demonstrates how serverless services can orchestrate AI‑powered media processing.