AI Agent Bake-Off Episode 1: Fashion AI Agent
A high-pressure 3-hour hackathon building a fashion AI agent for closet recording and virtual try-ons using Google Cloud’s Agent Development Kit, Gemini, Imagen, and Veo.
In this episode, we tackled the “imagination gap” in online shopping—helping people visualize how clothes will actually look on them. The agent includes closet cataloging, virtual try-on capabilities, and smart suggestions for unworn items in your wardrobe.
Watch the Full Episode
Key Technologies Used
- Gemini - Google’s flagship multimodal model for natural language understanding
- Agent Development Kit (ADK) - Framework for orchestrating multi-agent workflows
- Imagen - Photorealistic image generation and manipulation
- Veo - Video generation capabilities
- Vertex AI - Underlying infrastructure
What We Built
Our Fashion AI Agent became a personal wardrobe concierge with three core capabilities:
- Closet Cataloging - Automatically catalog wardrobe items with AI vision
- Virtual Try-On - See how different clothing combinations look on you
- Unworn Clothes Tracker - Get suggestions for neglected items in your closet
For a detailed write-up of this project, including the multi-agent architecture patterns we used, check out the full blog post.