06e8809c2a
* feat(updates): notify when a newer GitHub release is available Check the configured release channel shortly after startup and every six hours. When an update exists, show a tray menu item and a green New badge on Settings → About, plus one desktop notification per version, so users can open release notes and update themselves. Co-authored-by: Jatin K Malik <jatinkrmalik@users.noreply.github.com> * refactor(updates): slim UpdateMonitor after ponytail review Drop unused APIs (tick/check_now/version injector/last_* state), integer-only GLib scheduling, and the local demo launcher. Keep tray notify-send local to avoid coupling to recognition_manager. Co-authored-by: Jatin K Malik <jatinkrmalik@users.noreply.github.com> * fix(updates): address Bugbot and harden notify persistence Persist last_notified_version only after notify-send spawns successfully, and sync the tray Update Available item from About-page checks without re-notifying. Also merge latest main and cover the new paths in tests. Co-authored-by: Jatin K Malik <jatinkrmalik@users.noreply.github.com> * test(tray): pin config mock for update-notification assertions CI on 3.9/3.10 showed empty MagicMock set call history after a successful notify-send spawn. Pin tray.config_manager to the test mock and record set() via side_effect so persistence checks are reliable across Python versions. Co-authored-by: Jatin K Malik <jatinkrmalik@users.noreply.github.com> * fix(updates): clear pending About update on failed lookup A failed About check hid the New badge but left _pending_update set, so clearing settings search could revive the badge while status still showed the failed lookup. Clear dialog pending state on failure (tray unchanged) and restore the badge via _set_about_update_badge. * fix(updates): ignore stale UpdateMonitor results after channel switch Bind the channel for each background check and discard the callback when the configured channel changed mid-flight, then re-check immediately — matching About-page stale-channel handling so the tray cannot show the wrong channel's update state. --------- Co-authored-by: Cursor Agent <cursoragent@cursor.com> Co-authored-by: Jatin K Malik <jatinkrmalik@users.noreply.github.com>
24 KiB
24 KiB
Updating Vocalinux
This guide explains how to update Vocalinux to the latest version.
What's New in v0.15.0
0.15.0 is a minor release on the stable line. It redesigns settings navigation, adds AppImage packages, expands the speech-language catalog (Hungarian and many more), cleans up continuous dictation spacing/capitalization, adds power/GPU controls, and improves Wayland IBus on compositors that ship ibus-wayland, on top of the 0.14 packaging work (Flatpak, AUR, configurable hotkeys).
Highlights
| Feature | Description |
|---|---|
| Searchable settings | Sidebar navigation with search replaces the seven-tab notebook (#601) |
| AppImage | Self-contained x86_64 and aarch64 builds attached to GitHub Releases (#573, #602) |
| Language catalog | ~33 selectable speech languages plus Auto-detect (incl. Hungarian); CLI choices derived from the same catalog; VOSK hides Whisper-only langs (#616, fixes #565) |
| Dictation polish | Auto-capitalize after . / ! / ?; append a trailing space after each completed utterance (#554, #608) |
| Auto-pause + keep-alive | Unload the model while configured apps run, or after an idle timeout (#592) |
| Vulkan GPU selection | Prefer a discrete GPU automatically; pick a device under Advanced settings (#590) |
| ibus-wayland | Prefer IBus on previously “unbridged” compositors when ibus-wayland is running (#614) |
CLI --version |
Print the installed version and exit (#563) |
New Features
- Searchable sidebar settings — Topic pages in a sidebar with live search instead of seven notebook tabs (#601)
- AppImage packaging — Relocatable x86_64 and aarch64 AppImages built in the release workflow (#573, #602)
- Expanded speech languages — Hungarian plus many high-demand Whisper languages in Settings/CLI; official Alphacephei VOSK models where available; Whisper-only languages stay hidden in the VOSK dropdown (#616 by @jatinkrmalik, fixes #565)
- Sidebar dictation controls — Recognition status, mic level, Test Dictation, and Close stay visible in the settings sidebar footer while switching pages (#618)
- Sentence capitalization — Capitalize at the start of dictation and after sentence-ending punctuation (#554, closes #553)
- Trailing space between utterances — Completed transcriptions leave a trailing space so the next session does not glue onto the previous sentence (#608, fixes #605)
- Auto-pause apps + model keep-alive — Optional unload while selected processes run; idle timeout unload for battery/Optimus laptops (#592, closes #445, #591)
- Vulkan discrete GPU auto-select + manual device — Prefer discrete devices; override in Advanced settings (#590, closes #589)
- IBus via ibus-wayland — On compositors previously treated as unbridged, use IBus when
ibus-waylandis available (#614 by @eiseleb47, closes #607) vocalinux --version— Print package version (#563, closes #555)
Bug Fixes
- Settings: Restore Custom Shortcut entry / Record / Set controls after the sidebar settings refactor (#619)
- Languages: Map English (India) (
en-in) to Whisper codeenfor whisper.cpp / Whisper / remote API (#617) - Audio: Stop Bluetooth mic probing from corrupting the heap — one PortAudio open per capture session, stop-before-close, no stereo probe on mono-only devices (#599, fixes #567)
- IBus: Keep engine teardown correct when parent
do_destroyfails (#613 by @eiseleb47, fixes #606) - Settings UI: Flatten info notices so helper text matches the rest of the dialog (#615)
- KDE Plasma Wayland: Skip unbridged IBus when
ibus-waylandis not present so injection does not silently fail (#577, fixes #574) - xdotool: Preserve input focus after injection (#564, fixes #549)
- Uninstall: Remove IBus data dir; stop the app by PID file; remove
~/.local/binlauncher wrappers reliably (#597, #569) - Installer: Prefer both libgirepository 1.0 and 2.0 when present (#583, fixes #571)
- AUR: Virtual
python-pywhispercppdependency; clipboard/wtype tools as optdepends (#579, #586) - UI: First-run dialog response without
Gtk.Dialog.do_response(#580, fixes #566) - Vosk: Italian and English-India entries in medium/large model tables (#551, fixes #550)
- Docs: Correct ydotool service setup guidance (#560, fixes #557)
Docs / website
- Marketing site redesign with workstation craft (#582)
- Languages marketing page with honest per-engine badges (#616)
- Multi-distro tray icon FAQ (#584)
- robots.txt no longer blocks indexable pages (#610)
- Website CI lint / action warning cleanup (#611)
See the full changelog.
What's New in v0.14.2
0.14.2 is a stability patch on the 0.14 series. The feature set is the same as 0.14.x; this release fixes IBus reliability and settings dialog sizing.
0.14 series highlights
| Feature | Description |
|---|---|
| Configurable Shortcuts | Bind any modifier combination to a key — e.g. Alt+R, Ctrl+Shift+V, or Super+F10 |
| FunASR / SenseVoice Remote API | Remote-API engine supports FunASR and SenseVoice via OpenAI-compatible endpoints |
| Flatpak packaging | Universal Flatpak (whisper.cpp) with sandbox-aware paths, global hotkeys, and Wayland paste injection |
| AUR package | Official Arch packaging and CI publish path |
| Layout-aware hotkeys | Combo keys respect non-US layouts |
| Wayland / IBus reliability | GNOME and KDE injection fixes, first-dictation FocusIn gate, engine process launch restored |
| Audio / hybrid-CPU | Recording device-index crash fixed; whisper.cpp no longer defaults to all cores on hybrid CPUs |
Bug fixes in v0.14.2
- IBus: Restore engine process launch after the Flatpak XDG path import change.
start_engine_processand the IBus component exec runibus_engine.pyby path, so the relative import failed withImportErrorand Vocalinux fell back to ydotool/clipboard paste (#534) - IBus: Wait for FocusIn before commit on scoped injection. Cold first activation on GNOME Wayland could commit before mutter bound a client context, so the first dictation of a session was dropped while logs still reported success (#533, fixes #523)
- Settings UI: Wrap each notebook tab in a vertical ScrolledWindow so the dialog fits 1080p monitors instead of growing past the screen; forward wheel events from unfocused combos/spins to the tab scroller and drop nested Advanced ScrolledWindow shadows (#538, #541)
See the full changelog.
What's New in v0.14.1
Highlights
| Feature | Description |
|---|---|
| Flatpak packaging | Universal Flatpak (whisper.cpp) with sandbox-aware paths, global hotkeys, and Wayland paste injection |
| AUR package | Official Arch packaging and CI publish path |
| Layout-aware hotkeys | Combo keys respect non-US layouts |
| Installer / injection fixes | sg on Ubuntu 26.04/Debian 13; XIM none treated as unset |
New Features
- Flatpak packaging — GNOME Platform 50, whisper.cpp + Vulkan, ydotool/wl-copy injection, evdev global shortcuts; build via
packaging/flatpak/(#484, closes #167) - AUR release package — PKGBUILD and CI publish for Arch Linux (#518)
Bug Fixes
- Hotkeys: Layout-aware combo keys for non-US layouts (#514)
- Installer:
sgnot found on Ubuntu 26.04 / Debian 13 (#524) - Text injection: Treat XIM
noneas unset (#512) - Web: Dependabot npm alerts in package-lock (#515)
Docs
- Refreshed v0.14 UI screenshots and website gallery (#521)
- README Star History and related polish
See the full changelog.
What's New in v0.14.0-beta
Highlights
| Feature | Description |
|---|---|
| Configurable Shortcuts | Bind any modifier combination to a key — e.g. Alt+R, Ctrl+Shift+V, or Super+F10 |
| FunASR / SenseVoice Remote API | Remote-API engine now supports FunASR and SenseVoice models via OpenAI-compatible endpoints |
| GNOME Wayland IBus Reliability | Text injection works again on GNOME Wayland with bare xkb layouts and engine restore fallbacks are fixed |
| Audio Crash Fix | Recording no longer crashes when the system audio device index changes between sessions |
| Hybrid-CPU Efficiency | whisper.cpp no longer defaults to all cores on hybrid Intel/AMD processors |
✨ New Features
- Configurable modifier+key hotkeys — The Settings dialog now lets you set custom shortcuts using any combination of Ctrl, Alt, Shift, and Super plus a letter/number key. The legacy defaults still work, and you can now bind combinations like
Alt+RorCtrl+Shift+V(#493) - Remote API FunASR/SenseVoice support — OpenAI-compatible remote endpoints can specify FunASR/SenseVoice model names (e.g.
sensevoice) and return richer response shapes; SenseVoice metadata labels are stripped before text injection (#468)
🐛 Bug Fixes
- GNOME Wayland/IBus: Restore text injection when only a bare
xkbengine is configured; the engine restore fallback now picks the correct IM engine instead of silently dropping text (#506, #500) - KDE Wayland/IBus: Restore the KDE Plasma Wayland IBus text-injection path that was regressed in recent compositor-detection changes (#502)
- Wayland injection: Wait for held modifiers (Ctrl/Alt/Shift/Super) to release before injecting text, preventing accidental shortcut triggers and garbled output on modifier-heavy workflows (#494)
- Shortcuts UI: Keep preset and custom shortcut selection exclusive — selecting a preset now clears the custom field, and setting a custom combo selects the "Custom Shortcut" preset (#509)
- whisper.cpp: Stop defaulting to all CPU cores on hybrid processors (Intel Performance + Efficient cores), which caused UI lag and excess battery drain (#492)
- Audio: Fix a crash on recording start when the selected audio device index no longer matches the current system enumeration (#499)
- Installer: Include
xselas a fallback for the Wayland clipboard path whenxclipis unavailable (#496)
🔧 Improvements
- Code style — Removed an outdated long comment about whisper.cpp default thread counts (#505)
See the full changelog.
What's New in v0.13.0-beta
🚀 Highlights
| Feature | Description |
|---|---|
| 🎙️ Guided Whisper Models | Pick a whisper.cpp size and specialization (English-only, quantized, Turbo) with in-app guidance |
| 🔌 Hotplug Keyboard Support | Shortcuts keep working on keyboards connected after startup |
| ✍️ Dictation Spacing | Spacing preserved between speech segments separated by a pause in the same session |
| 🖥️ Wayland Reliability | Fixes silent text drops on wlroots/COSMIC compositors and garbled non-US-layout output |
✨ New Features
- Guided whisper.cpp model variants — The Settings dialog now splits whisper.cpp selection into Model Size and Specialization, exposing English-only, quantized (Q5/Q8), Large v3 Turbo, and legacy large models with language-aware recommendations and hover guidance. Exact model IDs (e.g.
medium.en-q5_0,large-v3-turbo) can also be passed to--model(#465)
🐛 Bug Fixes
- Dictation: Preserve spacing between speech segments separated by a pause (#464)
- Shortcuts: Rescan for hotplugged keyboards so shortcuts work on devices connected after startup (#467)
- KDE Plasma Wayland: Detect KDE Plasma Wayland sessions and guide you to enable IBus Wayland when
wtypeinjection fails (#466) - Wayland: Fix garbled output on non-US keyboard layouts and a clipboard-copy hang; ydotool now pastes through the clipboard (#480)
- Wayland/IBus: Use wtype/ydotool instead of IBus on compositors that don't bridge IBus to native apps like COSMIC, Sway, and Hyprland (#486)
- Wayland/IBus: Require a real IM engine on Wayland so a bare
xkblayout no longer causes silent text drops on GNOME/Mutter and similar (#478) - Wayland: Preserve the keyboard layout on Wayland by not running
setxkbmap(was flipping XWayland apps tous) (#474) - UI: Cap the settings dialog height on high-resolution displays (#465)
🔧 Improvements
- Performance: Faster ydotool text injection via an explicit
--key-delay(#488) - Website: New documentation pages for Remote API, Silero VAD, advanced whisper.cpp settings, and desktop reliability (#470)
- CI: Automatic pull-request labeling by changed files (#473)
See the full changelog.
What's New in v0.12.0-beta
🚀 Highlights
| Feature | Description |
|---|---|
| 🌐 Remote API Engine | New backend for compatible remote transcription services |
| 🎙️ Silero VAD | Neural VAD drops silence-only buffers for cleaner dictation |
| 🧵 Thread Safety | Hardened Remote API, IBus, and text injection threading behavior |
| 🔌 IBus Reliability | Preserves user engines for dead keys and scoped activation |
| ⚙️ Settings Polish | Advanced-only Remote Server controls and lower dialog height |
| 📦 Installer & Models | CUDA auto-remediation and corrected model download metadata |
✨ New Features
- Remote API speech recognition engine — Configure compatible remote transcription services alongside local engines (#335)
- Silero VAD — Neural voice activity detection filters silence-only buffers when ONNX Runtime support is installed (#447)
🐛 Bug Fixes
- Threading: Harden Remote API, IBus, and text injection thread safety (#452)
- IBus: Preserve user engines for dead keys and capture the current engine during scoped activation (#457, #458)
- UI: Keep the Remote Server section behind the Advanced toggle and reduce settings dialog height (#454, #456)
- Installer: Harden CUDA diagnostics with auto-remediation and behavioral tests (#451)
- Models: Correct whisper.cpp and VOSK download size metadata (#453)
- Startup: Allow launch without the pynput backend (#448)
- Website: Clarify speech demo browser support (#449)
🔧 Improvements
- Developer docs — Remote API test server instructions for backend testing (#455)
- Community — GitHub Sponsors funding configuration added
- Behavioral coverage — CUDA diagnostics and release-facing reliability fixes include targeted tests
See the full changelog.
What's New in v0.10.2-beta
🚀 Highlights
| Feature | Description |
|---|---|
| 🌍 Non-ASCII Text Injection | ydotool now falls back to clipboard paste for non-ASCII characters (á, é, ñ, etc.) |
| 🔌 IBus on Wayland | IBus now detected and started correctly on Wayland without legacy env vars |
| 🚀 IBus Engine Startup | Engine process now starts before registration check — fixes startup on some systems |
| 📦 Pop!_OS / Ubuntu 24.04+ | Added missing system dependencies (cmake, libcairo2-dev, libgirepository1.0-dev) |
| ⚡ Code Quality | Systematic refactor across 20 quality dimensions |
| 🖼️ Website OG Image | Redesigned Open Graph image — cleaner and more professional |
🐛 Bug Fixes
- #362 / #376: Handle non-ASCII characters with ydotool via clipboard paste fallback
- #360 / #361: Start IBus engine process before checking registration
- #381: Detect IBus on Wayland without legacy env vars and fix text injection
- #379: Add missing dependencies for Pop!_OS and Ubuntu 24.04+
🔧 Improvements
- Systematic code quality refactor across 20 dimensions (#377)
- Clarify missing GNOME AppIndicator support on Debian (#385)
- Redesigned OG image for vocalinux.com (#392)
See the full changelog.
What's New in v0.10.1-beta
🚀 Highlights
| Feature | Description |
|---|---|
| 🖼️ Tray Resource Reliability | Bundled package resources now prevent missing system tray icons |
| 🧠 Engine Switch Safety | Recognition now stops before engine changes to avoid segfaults |
| ⌨️ Keyboard Layout Preservation | XKB layout is preserved when activating the Vocalinux IBus engine |
| 🪟 Settings Dialog Compatibility | Added an explicit Close button for improved WM interoperability |
| ⚡ Suspend/Resume Recovery | App now automatically recovers speech recognition and keyboard shortcuts after system suspend/resume |
| 🎤 Push-to-Talk Reliability | Fixed premature transcription triggering on silence during push-to-talk mode |
✨ Scope
- Patch-focused release — No new feature surface; this version is dedicated to stability and compatibility fixes
- Desktop reliability hardening — Improved behavior across tray, engine switching, settings dialog actions, and keyboard layout handling
- Suspend/Resume stability — New D-Bus handler ensures app survives system sleep cycles
🐛 Bug Fixes
- #349 / #354: Bundle resources in package to fix missing system tray icons
- #350 / #355: Stop recognition before switching engines to prevent segfaults
- #323 / #356: Add Close button to settings dialog for WM compatibility
- #292 / #343: Preserve XKB layout when activating Vocalinux IBus engine
- #359: Prevent premature transcription during push-to-talk silence
- #367 / #369: Auto-recover speech recognition after system resume via new suspend handler
- #371: Restart keyboard shortcut backend after system resume
- #372: Delay keyboard restart to allow USB device re-enumeration after resume
🔧 Improvements
- Bumped npm/yarn dependency group across the web workspace (#346)
- Bumped
brace-expansionin development dependencies (#357) - Disabled copy-to-clipboard by default in Settings (#370)
See the full changelog.
What's New in v0.9.0-beta
🚀 Highlights
| Feature | Description |
|---|---|
| ⌨️ Left/Right Modifier Keys | Choose Left Ctrl vs Right Ctrl (etc.) as your shortcut trigger |
| 🔔 Sound Effects Toggle | Enable or disable audio feedback from the Settings dialog |
| 📋 Wayland Clipboard Fallback | Automatic clipboard copy when virtual keyboard injection isn't available |
| 🛠️ Installation Polish | Better pipx/Debian guidance and headless display detection |
✨ New Features
- Left/Right Modifier Key Distinction — Shortcuts now support
Left Ctrl,Right Alt, etc., with grouped UI in Settings - Sound Effects Toggle — New Audio Settings toggle to silence start/stop/error sounds
- Clipboard Fallback for Wayland — Auto-copies text via
wl-copy/xclipwhen injection unavailable (KDE Plasma etc.) - Display Availability Check — Graceful error message when running in headless environments
🐛 Bug Fixes
- #308: Distinguish left vs right modifier keys (evdev + pynput backends)
- #307: Remove unwanted leading space when starting a new transcription session
- #305: Pass configured shortcut mode to
KeyboardShortcutManageron startup - #299: Add clipboard fallback for Wayland compositors without virtual keyboard support
- #289: Improve Debian/pipx installation error messages and cross-distro dependency guidance
🔧 Improvements
- Grouped shortcut selector — Settings dropdown now organises shortcuts by Either/Left/Right side
- pipx documentation — New
DISTRO_COMPATIBILITY.mdsection for pipx users
See the full changelog.
What's New in v0.8.0-beta
🚀 Highlights
| Feature | Description |
|---|---|
| 🎤 Push-to-Talk Mode | Hold the shortcut key to speak, release to stop |
| ⚙️ Voice Commands Toggle | Enable or disable voice commands, with VOSK auto-enable in auto mode |
| ⌨️ Shortcut Reliability | Improved callback lifecycle and mode switching stability |
| 🧠 Input Compatibility | Better IBus detection and audio device/channel compatibility |
✨ New Features
- Push-to-Talk Shortcut Mode — Added hold-to-speak mode alongside double-tap toggle mode
- Mode-aware Shortcut UI — Updated settings text and behavior for toggle vs push-to-talk workflows
- Voice Commands Optional — Voice commands can be disabled, with automatic enable behavior for VOSK
🐛 Bug Fixes
- #277: Detect active IBus input method before using IBus injection
- #275: Detect and use device-supported channel count
- #268: Prevent GTK startup dialog crash on Fedora
- #261: Resolve text injection issues for better reliability
- #259: Prevent recognition thread state flicker
- #262/#263: Auto-detect audio sample rate for better hardware compatibility
🔧 Improvements
- Web SEO Enhancements — Added 8 additional optimized pages for discoverability
- Homepage Refresh — Updated voice-themed visual polish on the web landing page
See the full changelog.
Quick Update
Vocalinux checks GitHub Releases in the background about every six hours (and shortly after startup). When a newer build is on your selected channel, the tray menu gains an Update Available… entry and Settings → About shows a green New badge; open that page for release notes and download links. Updates are not installed automatically — re-run the installer (or your package manager) as below.
If You Installed via curl (Recommended)
Simply re-run the installation command:
curl -fsSL https://raw.githubusercontent.com/jatinkrmalik/vocalinux/main/install.sh | bash
The installer will:
- ✅ Detect and stop any running Vocalinux processes
- ✅ Update your existing installation
- ✅ Preserve your configuration and models
- ✅ Install any new dependencies
- ✅ Attempt to install neural VAD support, falling back safely if ONNX Runtime is unavailable
If You Installed from Source
cd vocalinux
git fetch origin
git checkout v0.15.0
./install.sh
Or to get the latest development version:
cd vocalinux
git pull origin main
./install.sh
Checking Your Current Version
python3 -c "import vocalinux; print(vocalinux.version.__version__)"
Troubleshooting
Update Issues?
If your update doesn't go smoothly, try a clean reinstall:
# Uninstall (keeps your config and models by default)
./uninstall.sh --keep-config --keep-data
# Or uninstall completely
./uninstall.sh
# Reinstall fresh
curl -fsSL https://raw.githubusercontent.com/jatinkrmalik/vocalinux/main/install.sh | bash
Old Version Still Running
# Prefer stopping from the tray, or kill by the PID in the lock file:
kill "$(tr -d '[:space:]' < "${XDG_DATA_HOME:-$HOME/.local/share}/vocalinux/instance.lock")"
# If you used IBus injection, also stop the engine:
kill "$(tr -d '[:space:]' < "${XDG_DATA_HOME:-$HOME/.local/share}/vocalinux-ibus/engine.pid")"
# Start fresh
vocalinux
Avoid pkill -f vocalinux — it matches any process whose command line
mentions the repo path (editors, shells, test runners) and can kill unrelated work.
Missing Dependencies
If you see dependency errors:
# Ubuntu/Debian
sudo apt update
sudo apt install -y python3-gi python3-gi-cairo gir1.2-gtk-3.0
# Fedora
sudo dnf install -y python3-gobject gtk3
# Arch
sudo pacman -S python-gobject gtk3