I’m a fan of the webcomic SMBC by Zach Wienersmith. It’s nerdy, consitently funny, and sometimes quite thought provoking. Often I remember a comic and want to show it to someone, but I’m unable to find it again. To solve this, I went ahead and wrote a script in Python to scrape the SMBC comics website and run optical character recognition (OCR) on each image using the easyocr package. The resulting text was put into a SQLite database and I wrote a little PHP code so that you can run a full-text search on it. The text has lots of errors (a lot of that is most certainly my fault, I didn’t try tuning the OCR parameters), but search is still quite effective.

The search page is available here: https://vikramkashyap.com/smbc_search.php