ThreadReads
Book recommendations from Hacker News — ranked by discussion quality
3,187 books indexed
Updated daily
About This Project

1 What is this?

ThreadReads indexes book recommendations from Hacker News discussions. Unlike simple mention-count lists, we rank books by the quality of discussion they generate — longer, deeper, more engaged conversations signal a book worth reading.

2 Methodology

We scrape Hacker News via the Algolia API, searching for book-related discussions across all time. An LLM (Claude) extracts book titles and authors from comment text. Each book is then scored using five weighted components:

Wt. Component Desc.
0.25 Mention Count log₂ scaled
0.20 Discussion Depth avg nesting
0.20 Comment Length log scaled
0.25 Diversity unique users + threads
0.10 Story Quality avg HN pts

Each component produces a 0–100 sub-score using logarithmic scaling (diminishing returns). The final score is a weighted sum.

3 Data Source

All data comes from the Hacker News Algolia API. The scraper runs daily via GitHub Actions and updates this site automatically. Book extraction uses Anthropic's Claude API.

4 Open Source

This project is open source. View on GitHub.