← Agents

Book Price Agent

In Progress

Finds the best price for any book across major retailers, used marketplaces, and library networks, including editions you didn't know existed.

bookspricessearchresearch

What it does

Books are priced inconsistently across the internet. The same title might cost $28 new, $4 used, nothing at a library, or be freely available as a legitimate PDF that the author posted on their website in 2009 and nobody told you about.

The Book Price Agent searches systematically across major retailers, secondhand marketplaces, and open-access repositories. It returns a ranked list of options sorted by price, with notes on condition, edition, and format.

It also flags when multiple editions exist and whether the differences matter. A 2003 edition of a statistics textbook is often identical to the 2019 one. A 2003 edition of a law textbook is not.

How to use it

Pass it a title, an author, or an ISBN. It handles the rest.

python agent.py "The Elements of Statistical Learning"
python agent.py --isbn 978-0387848570
python agent.py "Gödel, Escher, Bach" --formats "paperback,ebook,pdf"

Output is a structured JSON file or a formatted markdown table, depending on your preference.

Examples

Query: "The Making of the Atomic Bomb" by Richard Rhodes

| Source | Format | Price | Condition | Notes | |--------|--------|-------|-----------|-------| | ThriftBooks | Paperback | $4.19 | Good | 1987 edition, minor shelf wear | | Amazon | Paperback | $19.99 | New | 2012 reprint | | Archive.org | PDF | Free | n/a | Scanned 1986 first edition | | Local library | Physical | Free | n/a | 2 copies, 1 available |

The agent noted that all editions share identical text; the price difference reflects print quality, not content.

Notes

Currently covers Amazon, ThriftBooks, eBay, AbeBooks, Open Library, Project Gutenberg, and standard library APIs for a handful of US cities. More sources are straightforward to add.

The agent does not scrape aggressively. It uses APIs and public interfaces, and respects rate limits. It is not designed for bulk queries.