{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiahncja2q4sgptv3iuigwyxdkmvjs7rr75vi6b67c2ygebhjnbaa4",
    "uri": "at://did:plc:34cg4tn4iwemk3v5k3n3adwf/app.bsky.feed.post/3mn5kor4rbr72"
  },
  "path": "/t/ocr-apps-that-actually-get-it-done/34352#post_18",
  "publishedAt": "2026-05-31T09:17:39.000Z",
  "site": "https://forum.f-droid.org",
  "textContent": "Hey,\n\nJust to let you know, I have tested the latest version of MakeACopy on F-Droid and it seems that the results are better than before indeed. I tested the same images/printed text as before and noticed less mistakes, I would say 99% accurate. Some things like 100% got interpreted as 1oo% but I really noticed an improvement. Suppose now it’s up to Paddle models to improve right?\n\nThough, regarding the app itself, I would add two suggestions if I may:\n\n  1. The cut preview should default to a basic square/rectangle. Reasoning: most people will usually try to photograph their documents semi-centered at least. Wins: Not only it would decrease the time the app would be taking trying to “guess” what parts of the image should and should not be covered (it does feel slow) it would also be less awkward for first usage to rely on the more usual method of merely adjusting height/width. Still, the users that DO NEED to trapezoid the cutting, have the option there and can use if they really see it fit. Personally, I find it harder to adjust the cut with the current method.\n\n  2. The first OCR preview could have a button for straight sharing the whole text. Reasoning: Sometimes you don’t want to save the content you just need to share it with another App or something. Wins: Right now I have to press the text, select all, share. Having a button for “share text preview” or something there would be a more diret approach.\n\n\n\n\nOther than that I really think the results are pretty good so far. Thank you!",
  "title": "OCR apps that actually get it done?"
}