AI · Book Discovery

BookSignal

Tell it what you want to learn — in plain English. BookSignal parses your intent, searches Google Books and Open Library, then ranks results with explainable AI reasoning. Not just titles. Reasons.

Live Demo
📚
What it does

Discovery that explains itself.

🧠

Intent parsing

Your casual query ("serious AI consciousness books, no shallow hype") becomes a structured search spec — topics, exclusions, depth, reading level — before a single API call is made.

🔍

Multi-source discovery

Searches Google Books and Open Library in parallel, then deduplicates and enriches results. Persistent caching means the second search is instant.

⚖️

Explainable ranking

Every recommendation comes with a reason — not just a score. Embedding similarity + LLM evaluation gives you a ranked list you can actually trust and push back on.

How it works

Three steps from query to shelf.

01
You describe what you want to learn
Natural language. No keywords, no filters. "Books on how organisations fail at AI adoption — practitioner-level, not academic."
02
Intent is parsed into a structured search spec
Claude extracts topics, desired depth, reading level, and what to exclude — then runs that spec against Google Books and Open Library simultaneously.
03
Results are ranked with reasons, not just scores
Vector embeddings measure semantic fit. An LLM evaluation pass adds a plain-English justification for each recommendation. You get a ranked list you can interrogate.
Tech stack
Claude API Google Books API Open Library Vector Embeddings Node.js Deployed on Vercel
Live preview

Try it here.

booksignal-sage.vercel.app