This post is my submission for DEV Education Track: Build Multi-Agent Systems with ADK.
What I Built
What I Built
GiftGenie is a multi‑agent system that transforms a simple description of a person into a curated list of thoughtful, budget‑friendly gift ideas. Instead of relying on one large prompt, the system uses four specialized agents—each deployed as its own Cloud Run microservice—to analyze the recipient, generate creative options, filter them by budget, and refine the final recommendations. This project demonstrates how distributed agents can collaborate to produce more reliable, structured, and high‑quality results than a single monolithic model.
Cloud Run Embed
cloudrun
https://your-cloud-run-frontend-url-here
Your Agents
Profile Analyzer
This agent interprets the user’s free‑form description of the gift recipient. It extracts traits, interests, and constraints and converts them into a structured profile. This step ensures the rest of the system works with clean, predictable data.
Key Learnings
The refinement step matters
The Refinement Agent had the biggest impact on perceived quality. Even when earlier agents produced good data, the final polish made the output feel intentional and user‑ready.
Top comments (0)