Skip to content

Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.

docs: drop TROUBLESHOOTING.md addition, moved into PR description instead

iSparsh61d995bb6434
diff --git a/TROUBLESHOOTING.md b/TROUBLESHOOTING.md
index 54183b85c..b45e8eb90 100644
--- a/TROUBLESHOOTING.md
+++ b/TROUBLESHOOTING.md
@@ -136,19 +136,6 @@ OpenWhispr tries clipboard methods in order: `wl-copy` (most reliable) → rende
4. Restart OpenWhispr and start meeting transcription again
5. No screen-share chooser is expected for Linux system audio; OpenWhispr captures the default sink monitor directly through PipeWire
-### Linux Auto-Learn Correction (AT-SPI2) Issues
-
-**Symptoms:** Auto-learn correction (Settings → Dictionary) never updates the custom dictionary after an edit, with no error shown, on Linux sessions other than GNOME.
-
-**Fix:**
-
-1. GTK/Gecko apps (Firefox, Claude Desktop, etc.) only register with AT-SPI2 when `org.gnome.desktop.interface toolkit-accessibility` is enabled — this is off by default on non-GNOME sessions (niri, Hyprland, sway, KDE). Enable it:
- ```
- gsettings set org.gnome.desktop.interface toolkit-accessibility true
- ```
-2. Restart the target app (Firefox, etc.) after enabling — the flag is read once at startup, not watched live.
-3. This affects only apps that don't expose their accessibility tree; apps that do keep working as before, and unsupported apps simply skip correction learning silently rather than erroring.
-
### Meeting Transcription Issues
**Symptoms:** Meeting detection not working, no transcription, audio not captured