fix(desktop): add precise local OCR reading

This commit is contained in:
Yun Chan 2026-11-21 10:15:49 +09:00
parent 59cdddbecb
commit 25fd3ac004
30 changed files with 626 additions and 15 deletions

View file

@ -12,6 +12,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- Additional Whisper model management UI
- Plugin system for custom pipelines
### Fixed
- Screen OCR reconstructs tightly spaced URL, email and punctuation tokens from their pixel positions instead of inserting a space between every recognized word. Ordinary word spacing is preserved; unrecognized glyphs are not guessed or rewritten.
- Add an on-demand precise-reading action to screen OCR. Only the selected crop goes to a loopback GLM-OCR model; the first action downloads about 2.7 GB, original text remains available, and failed, truncated, repeated or cancelled results never replace it.
## [1.18.4] - 2026-11-20