for my uses I just use the apis of google lens and bing ocr, they're not publicly open sourced but they can be reverse eng. easily , I think there are github repos that use them,, the good thing is that they don't need tokens and have no limit per ip ( not that i did face anyways), also they give bounding rects coords,, but unfortunately couldn't find public apis for parsing tables, they just give coords of lines and words and their texts, a bit dirty but good enough for me
Tables are pretty accessible as via both DocumentApp with Apps Script and Docs API. You can also fetch an Export link that will produce a Docx that you can then extract the tables from using an external API. Document AI also does a pretty good job of table extraction if you want to stay in the ecosystem.
for my uses I just use the apis of google lens and bing ocr, they're not publicly open sourced but they can be reverse eng. easily , I think there are github repos that use them,, the good thing is that they don't need tokens and have no limit per ip ( not that i did face anyways), also they give bounding rects coords,, but unfortunately couldn't find public apis for parsing tables, they just give coords of lines and words and their texts, a bit dirty but good enough for me
Tables are pretty accessible as via both DocumentApp with Apps Script and Docs API. You can also fetch an Export link that will produce a Docx that you can then extract the tables from using an external API. Document AI also does a pretty good job of table extraction if you want to stay in the ecosystem.