BookSpot AI: Vision Book Recommender

Learn how to combine OCR, CLIP image‑text models, and an LLM to detect books in photos and generate personalized recommendation lists.

Overview

I’m building a computer vision–powered web app that helps people discover the right books from physical stalls or secondhand shops. You take a photo of random books, and the AI identifies them and recommends which ones best match your interest (e.g., health, AI, self-help). It uses OCR for title detection, CLIP for image–text understanding, and an LLM for semantic book recommendations.

Tech stack