Search Errors with Magic Books

Platform decay came for the Magic Book, but, fortunately I found some wet wipes.

The Magic Books searched for information and cover images from two sources: Open Library (part of the non-profit Internet Archive) and what I described as a "private non-commercial source," which was really some old credentials I had for some university library systems as a result of work for them I did years ago.

Back when I created the Magic Book (2017!), the Open Library's search results weren't all that great for a general purpose tool, so I asked the folks at the university system if I could use my login to supplement it. They said it would be fine as long as I wasn't reselling content or serving the content of any works, just conducting searches and occasionally grabbing a (fair use) cover image. Oh, and the traffic couldn't get out of hand. (The traffic never got out of hand: they never even noticed.)

However, academia is kind of like everything else—largely at the mercy of Big Tech—and a series of retirements, redundancies, and administrative reshufflings meant my credentials to access those systems were terminated at some point in late May 2026.

So: I've rejiggered the Magic Book so that it now relies solely on Open Library for search and cover images. Ironically, this makes the back end simpler. Open Library has come a long way since 2017 and it's now much more functional as a general purpose tool.

However, Open Library is not without its own conflicts and controversies, and—of course—if Big Tech has their way public libraries of any sort will be eliminated. So I can't guarantee this is a permanent or long-term solution, but it should work for a while.

As for the completeness of Open Library's data and search accuracy, ISBN searches remain pretty functional in most cases. Searching for titles remains…uh, I'll go with whimsical.