2026
Natural-language search engine for computer peripherals
A FastAPI service that takes natural-language queries and returns the most relevant peripherals from a catalog, exposed through a Cloudflare tunnel.
Problem
Find the right peripheral (keyboard, mouse, headset) by describing it in your own words, without rigid filters.
Data
A hand-built catalog of 75 peripherals (mice, keyboards, headsets, etc.) with 21 attributes: brand, model, description, connection, DPI, weight, battery, microphone, price in ARS and rating.
Method
Catalog normalization, text cleanup with NLTK stopwords, vectorization and similarity ranking with scikit-learn, and a REST API with CORS.
Result
A working API that answers free-form queries over the full peripherals catalog.
Stack
- Python
- FastAPI
- NLTK
- scikit-learn
- pandas
- Cloudflare Tunnel
Context
Course: Natural Language Processing · 2026