Urgent.News

What's breaking now, across thousands of outlets.

Editions

AI

Multi-Agent Gift Recommendation Engine Powered by Google ADK & Gemini

This post is my submission for DEV Education Track: Build Multi-Agent Systems with ADK . Finding the perfect, thoughtful gift shouldn't feel like a chore. Whether it's for a birthday, anniversary, or holiday, we all experience gift-buying paralysis: Generic suggestions : "Just buy them a mug or a generic gift card." Budget anxiety : Falling in love with an idea only to find out it costs 3x what…

GiftAdvisor is a sophisticated gift recommendation system that utilizes Google's Agent Development Kit (ADK) and the Gemini model to provide personalized gift suggestions based on a recipient's profile. The system is built with three specialized AI agents that work together in a multi-agent pipeline.

The Profile Analyzer Agent is responsible for extracting core interests, lifestyle dimensions, and anti-preferences from the user's input. This agent uses the LlmAgent component from ADK and the Gemini model to understand the recipient's description, occasion, and relationship, then saves this structured analysis in session state.

Next, the Idea Finder Agent generates 6-10 candidate gift ideas across various categories, such as experiential, practical, consumable, and sentimental items. The agent also attaches realistic estimated market prices for each idea and saves the candidate gifts in session state.

Finally, the Budget Filter Agent evaluates the candidate gifts against the user's budget limit. It eliminates any items that exceed the budget and replaces them with budget-friendly alternatives. The agent then ranks the final recommendations into budget tiers and provides specific buying advice. The final report is saved in session state.

The entire system is deployed on Google Cloud Run, which allows for cost-effective scaling to zero instances when idle. The system uses a modern Glassmorphism UI, with a dark mode interface, preset profiles, an interactive budget slider, and live pipeline stage tracking. Additionally, users can export the recommendations in various formats, including Markdown, JSON, Clipboard, or PDF.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

Claude Prompt Caching: Why Agent Loops Miss the 20-Block Lookback

Your agent starts a run with cache_read_input_tokens at 40K and climbing. Twelve tool calls later, reads drop to zero and cache_creation_input_tokens jumps to the full conversation length — on every…

  • Claude Prompt Caching limited to 20 blocks
  • Agent loops exceed 20-block lookback window
  • Breakpoints spaced 15 blocks apart fix issue

More from Friday 21 August →