What it does

GenMedia Creative Studio is a browser-based demo app that exposes Google Cloud’s generative media APIs—image generation (including the whimsically named “Nano Banana” models), video via Veo, music via Lyria, and speech via Chirp 3 and Gemini TTS—through a single interface. It also bundles pre-built workflows like “Shop the Look” and “Interior Designer” that chain these models together. The whole thing is built on Mesop, an internal-flavored Python UI framework Google uses for quick AI prototypes.

The interesting bit

The experiments/ folder includes MCP (Model Context Protocol) servers, which means this isn’t just a pretty UI—it’s also exploring how generative media tools can plug into broader agent workflows. That’s the part worth watching if you care about composable AI systems rather than one-off generation.

Key highlights

• Covers the full generative media stack: images, video, music, speech
• Pre-built workflows for character consistency, virtual try-on, moodboards
• Deploys via Terraform + Cloud Run, with IAP and custom domain support
• Includes MCP servers and standalone experiments for advanced use cases
• One-click Cloud Shell launch if you don’t want to set up locally

Caveats

• Explicitly not an officially supported Google product; demo-only, not production-grade
• Chrome recommended; Safari and Firefox may break on some features
• README is vague on whether API costs are your own (they almost certainly are)

Verdict

Worth a spin if you’re already on Google Cloud and want to survey what Vertex AI’s generative APIs can do without writing boilerplate. Skip it if you need a production-ready creative suite or if you’re not prepared to foot the API bill.