Quoril
Quoril is an AI-powered buyer agent for online marketplaces. It combines intelligent search with predictive pricing to help users find the best deals across platforms, acting as an autonomous purchasing assistant that understands context, preferences, and market trends.
Motivation
Online marketplaces are flooded with listings. Finding the right item at the right price requires sifting through duplicates, comparing across platforms, and timing purchases well. I wanted to build a system that could handle all of that autonomously.
How it works
- Natural language search powered by LLMs to understand user intent beyond keyword matching
- Cross-platform aggregation to surface relevant listings from multiple marketplaces
- Predictive pricing models that estimate fair value and flag good deals
- Agent-based architecture where buyer agents negotiate and execute purchases on behalf of users
Technical details
The frontend is built in React with a clean conversational interface. The backend runs on Express.js, orchestrating multiple LLM calls for query understanding, listing evaluation, and price prediction. The system uses a retrieval-augmented pipeline to keep results grounded in real marketplace data.