Show HN: Diff filtering, text mapping, and windowed transforms for LLM apps (opens in new tab)

(github.com)

3 pointszalzal1y ago3 comments

I couldn't find good Python libs to do this so ended up writing it myself.

3 comments

3 comments · 2 top-level

beebaween1y ago· 1 in thread

Can this be used with gemini 2 flash while OCR'ing pdfs?

It doesn't deal with images at all, but depending on what you're doing it might be useful in cleaning up the OCR text, e.g. asking LLMs to fix errors and then filtering the diffs.

zalzalOP1y ago

I'm curious if it's useful! How do others try to solve use cases like this of mapping data between docs or strictly controlling LLM edits?

j / k navigate · click thread line to collapse