Skip to content

v0.4.5

Choose a tag to compare

@yakuda-stack yakuda-stack released this 28 Sep 17:12
· 3 commits to main since this release

[v0.4.5] – 2026-09-28

  • πŸ” Lens (live translation): the type trigger now works in automatic mode too – πŸͺ„ Auto ↔ πŸ” Lens (manual: πŸ–Ό β†’ πŸ“ β†’ πŸ”³ β†’ πŸ” Lens). With πŸ” the shutter takes no photo but photographs the same area again every 1–10 s (Options β†’ Shot; blue frame in VR, stays in front of your head) and the app translates it – fastest with LibreTranslate. Pictures go to live/live.png, not the gallery; unchanged text isn't translated again. Open a new frame to stop. New icons πŸͺ„ Auto and πŸ” Lens. Needs a layer rebuild.
  • AI answers appear live while Claude / Gemini / a custom command (e.g. Ollama) is still writing (ChatGPT/Codex usually delivers the answer in one piece)
  • ☁/πŸ”’ privacy note for every service (Options β†’ Translation + tooltip in the Main dropdown): whether the recognised text is sent to the internet – only text, never the photo
  • Tests: layer.json defaults in Python and Rust are compared automatically; new GitHub Actions workflow runs cargo test, clippy, pytest and the smoke test on every push
  • Code: main_page.py split into main_translation.py, main_history.py, main_install.py, main_live.py (Lens) (+ ui/panel_window.py)