Release History

Vocalinux Changelog

Track every release and see how Vocalinux has evolved. From initial alpha to stable releases, follow the journey of Linux voice dictation.

v0.16.2

StableLatest2026-09-05

KDE: skip leftover IBus when it is not the session IM so dictation types into Kate, browsers, and terminals (PR #753, fixes #752)
Wayland shortcuts: wtype and ydotool deliver real chords instead of refusing or silently doing nothing (PR #715)
IBus shortcuts: route X11_IBUS through xdotool and WAYLAND_IBUS through wtype/ydotool (PR #716)
Voice command delete that: send real BackSpace key events instead of U+0008 text (PR #714)
Installer: Fedora and Arch need glslc/shaderc packages, not glslang (PR #763, #604)
Nightly: stamp version.py before python -m build so wheel metadata matches the filename (PR #762)
Release integrity: checksums, signatures, and pinned builders for release artifacts (PR #759, epic #701 phase 5)
CI: gate AUR PKGBUILD builds on every PR; test the distros the docs promise and fix docs drift (PR #772, #773)
Site: VocaGateway family card is Beta; README logo, badges, and privacy copy (PR #765, #764)
Deps: bump the github-actions group (PR #766)

Remote API guide Silero VAD guide Advanced settings Reliability overview

View release on GitHub

v0.16.1

Stable2026-08-30

Startup uses the engine's own model size instead of a leftover generic key; the model is saved only after it loaded (PR #684, #685, fixes #681)
Unused leftover-model list shows every row; one ConfigManager so Settings writes stick (PR #686, #691, fixes #683, #689)
Tray stays idle after leftover transcription on toggle stop, and the missing-model notification can download the recommended model (PR #741, #687)
Terminals get Ctrl+Shift+V paste; GNOME XWayland layout is restored after scoped IBus inject (PR #734, #742)
Installer requires Python 3.11, verifies model downloads, and no longer invents GPUs (PR #713, #736)
AppImage built against a glibc floor that boots on Debian 12 through current Fedora (PR #743, #744)
AUR: build against Arch extra setuptools 84, and skip context_params on AUR pywhispercpp 1.4 so the app starts
Settings dropdowns share a width; About is quieter with family platform marks (PR #754)

View release on GitHub

v0.16.0

Stable2026-08-23

In-app update checker (stable/nightly) plus tray Update Available when GitHub has a newer release (PR #631, #645)
New installs default to hold Right Alt push-to-talk; existing configs keep their shortcut (PR #648)
Searchable language list and delete unused downloaded speech models (PR #672, #671)
Family dictation tone picker in Settings → Audio: Lift, Flick, Ember, Step, Voca, Soft, Chirp, Scale, Drop, Glass, Off, plus Preview. New installs default to Voca. Catalog tones are the family preview WAVs, not synthesized files (PR #707, #708)
License migrated to AGPL-3.0 (PR #660)
App icon, tray states, and site favicons use the shared Voca family mic instead of the old Linux rounded-rect (PR #704)
Installer hardened with Justfile and uv lockfiles; distro python3-gi required (no pip sdist of PyGObject). Epic #701 remains open (PR #700, #705, #706)
Test Dictation no longer reports no speech when recognition never started (PR #702)
IBus/X11 reliability: restorable scoped injection, engine restore after teardown, XKB layout restore (PR #623, #643, #665)
Audio, clipboard, and AppImage/GPU packaging fixes (PR #629, #673, #588, #646, #674, #637)
Settings About page groups this app, the VocaHQ family, and talk-to-us links (PR #718)
vocalinux.com restyled to the Voca family workbench; Discord and X links point at VocaHQ (PR #728, #729, #722, #717)

View release on GitHub

v0.15.0

Stable2026-07-28

Searchable sidebar settings with live search replace the seven-tab notebook (PR #601)
AppImage packages for x86_64 and aarch64 on GitHub Releases (PR #573, #602)
Expanded speech-language catalog (~33 languages + Auto-detect), including Hungarian; VOSK only lists languages with official models (PR #616, fixes #565)
Settings: dictation status / mic level / Test Dictation / Close live in the sidebar footer (PR #618)
Settings: Custom Shortcut Record/Set controls show again (PR #619)
Languages: English (India) maps to Whisper code en for whisper.cpp / Whisper / remote API (PR #617)
Auto-capitalize after sentence punctuation; trailing space so the next utterance does not glue on (PR #554, #608)
Auto-pause competing apps and idle model keep-alive unload for battery/GPU headroom (PR #592, closes #445, #591)
Vulkan: auto-select discrete GPU and pick a device in Advanced settings (PR #590, closes #589)
Wayland: use IBus on previously unbridged compositors when ibus-wayland is running (PR #614, closes #607)
IBus: keep engine teardown correct when parent destroy fails (PR #613, fixes #606)
CLI: vocalinux --version (PR #563, closes #555)
Settings info notices flattened; Bluetooth mic heap-corruption fix; KDE unbridged-IBus skip when ibus-wayland is absent; xdotool focus preserve; installer/AUR fixes (PR #615, #599, #577, #564, #583, #569, #597, #579, #586)
Marketing site redesign; languages page per-engine badges; robots.txt indexing fix (PR #582, #616, #610)

View release on GitHub

v0.14.2

Stable2026-07-17

IBus: restore engine process launch after Flatpak XDG path import so the engine no longer dies with ImportError and falls back to ydotool/clipboard paste (PR #534)
IBus: wait for FocusIn before commit on scoped injection so the first dictation of a session is not dropped on GNOME Wayland (PR #533, fixes #523)
Settings UI: notebook tabs scroll so the dialog fits the monitor; wheel events from unfocused combos/spins reach the tab scroller (PR #538, #541)

View release on GitHub

v0.14.1

Stable2026-07-17

Flatpak packaging for universal distribution: whisper.cpp engine, XDG sandbox paths, global hotkeys via evdev, Wayland text injection via wl-copy + ydotool (PR #484, closes #167)
AUR package and CI publish path for Arch Linux (PR #518)
Layout-aware combo keys so custom shortcuts work on non-US keyboard layouts (PR #514)
Installer fix for sg not found on Ubuntu 26.04 / Debian 13 (PR #524)
Text injection treats XIM none as unset (PR #512)
Website screenshot gallery refresh and Dependabot npm alert fixes (PR #521, #515)

View release on GitHub

v0.14.0-beta

Beta2026-07-13

Configurable modifier+key hotkeys: set custom shortcuts with any combination of Ctrl, Alt, Shift, and Super plus a letter/number key, e.g. Alt+R or Ctrl+Shift+V (PR #493)
Remote API engine now supports FunASR/SenseVoice models via OpenAI-compatible endpoints; SenseVoice metadata labels are stripped before text injection (PR #468)
GNOME Wayland IBus text injection restored when only a bare xkb engine is configured; engine restore fallback now picks the correct IM engine (PR #506, #500)
KDE Plasma Wayland IBus text-injection path restored after recent compositor-detection regressions (PR #502)
Wayland text injection now waits for held modifiers to release before typing, preventing accidental shortcut triggers and garbled output (PR #494)
Shortcuts UI keeps preset and custom shortcut selection exclusive: selecting a preset clears the custom field, and setting a custom combo selects the Custom Shortcut preset (PR #509)
whisper.cpp no longer defaults to all CPU cores on hybrid processors, improving UI responsiveness and battery life (PR #492)
Fixed a crash on recording start when the selected audio device index no longer matches the current system enumeration (PR #499)
Installer includes xsel as a fallback for the Wayland clipboard path when xclip is unavailable (PR #496)
Removed an outdated long comment about whisper.cpp default thread counts (PR #505)

View release on GitHub

v0.13.0-beta

Beta2026-06-30

Guided whisper.cpp model selection: pick a size plus a specialization (English-only, quantized Q5/Q8, or Large v3 Turbo) through split Model Size and Specialization dropdowns with in-app guidance; the --model flag also accepts exact IDs like medium.en-q5_0 and large-v3-turbo (PR #465)
Dictation now keeps a space between segments spoken with a pause in between, so words no longer run together after a silence (PR #464)
Keyboard shortcuts now work on keyboards hotplugged after Vocalinux starts, with the evdev backend rescanning for new devices and recovering from disconnects (PR #467)
Vocalinux now detects KDE Plasma Wayland sessions and points you to enable IBus Wayland for reliable text injection, surfaced during install and when wtype injection fails (PR #466)
Wayland: fixed garbled text on non-US keyboard layouts (AZERTY/QWERTZ/Dvorak) and a clipboard-copy hang; ydotool now pastes through the clipboard, which is layout-independent (PR #480)
Wayland: use wtype/ydotool instead of IBus on compositors that don't bridge it to native apps like COSMIC, Sway, and Hyprland, fixing silent text drops (PR #486)
Wayland/IBus: require a real IM engine before using IBus, so a bare xkb layout no longer causes silent text drops on GNOME/Mutter and other compositors (#478)
Wayland: keep the keyboard layout intact by not running setxkbmap, which was flipping XWayland apps to us after dictation (#474)
Faster ydotool text injection via an explicit --key-delay (PR #488)
Settings dialog height capped on high-resolution displays (PR #465)
Refreshed website docs with new feature pages for Remote API, Silero VAD, advanced whisper.cpp settings, and desktop reliability, plus responsive layout polish (PR #470)

View release on GitHub

v0.12.0-beta

Beta2026-06-07

Remote API speech recognition engine with installation and configuration support (PR #335)
Silero VAD drops silence-only buffers for cleaner dictation when ONNX Runtime support is available (PR #447)
Thread safety hardening for Remote API, IBus, and text injection paths (PR #452)
IBus preserves user engines for dead keys and captures the current engine during scoped activation (PR #457, #458)
Remote Server settings now respect the Advanced toggle and the settings dialog fits lower-resolution screens (PR #454, #456)
CUDA diagnostics now include auto-remediation and behavioral tests (PR #451)
Corrected whisper.cpp and VOSK model download size metadata (PR #453)
Startup now works without the pynput backend (PR #448)
Remote API developer test server documentation (PR #455)
Website speech demo browser support clarification (PR #449) and GitHub Sponsors funding configuration

View release on GitHub

v0.11.0-beta

Beta2026-05-30

New Advanced Settings tab with whisper.cpp anti-hallucination parameters - temperature, no_speech_threshold, max segment length, and more (PR #415)
IBus engine readiness probe at startup with hardened retries (PR #391)
IBus runtime failure recovery without app restart (PR #411)
IBus engine instance destruction handled on keyboard layout switch (fixes #388, closes #389)
Preserve final speech on stop - no more truncated transcriptions (fixes #401)
Play stop sound immediately on release and after audio thread joins (PR #426, #436)
Repair pywhispercpp library loading in installer (PR #433)
Reduce whisper.cpp CPU threads and ensure GPU backend builds in dev mode (PR #439)
Correct openSUSE Tumbleweed dependencies with fallback handling (PR #418, #420)
Harden Debian compatibility layer in installer (PR #437)
Add Python 3.14 support and bump lxml>=6.1.0 (fixes #404)
Validate pyproject.toml/setup.py content before entering local repo mode (fixes #396)
Reuse existing whispercpp builds during install (PR #421)
Refresh ldconfig after openSUSE typelib install, clarify python3XY placeholder convention (PR #438)
Clean up runtime log noise and cache hardware detection
Test coverage: recognition internals, IBus edge cases, CI notification suppression (PR #410, #414)
Clarify PyPI installation requirements (PR #423)
Dependency bumps: Next.js security updates (PR #399, #429), PostCSS

View release on GitHub

v0.10.2-beta

Beta2026-04-08

Handle non-ASCII characters (á, é, ñ, etc.) with ydotool via clipboard paste fallback (fixes #362, PR #376)
Detect IBus on Wayland without legacy env vars and fix text injection (PR #381)
Start IBus engine process before checking registration to fix startup on some systems (fixes #360, PR #361)
Add missing dependencies for Pop!_OS and Ubuntu 24.04+ including cmake, libcairo2-dev, libgirepository (PR #379)
Systematic code quality refactor across 20 dimensions (PR #377)
Clarify missing GNOME AppIndicator support on Debian (PR #385)
Redesigned OG image for vocalinux.com - cleaner, professional, text-based layout (PR #392)
Test coverage improvements: mock Notify module, tray degraded-startup, IBus socket-readiness branches (PR #384, #386, #390)

View release on GitHub

v0.10.1-beta

Beta2026-03-30

Bundled package resources to prevent missing system tray icons (fixes #349, PR #354)
Stopped recognition before engine switches to prevent segfaults (fixes #350, PR #355)
Added a dedicated Close button in Settings for better WM compatibility (fixes #323, PR #356)
Preserved XKB layout state during Vocalinux IBus activation (fixes #292, PR #343)
Auto-recover speech recognition after system suspend/resume via new D-Bus handler (fixes #367, PR #369)
Restart keyboard shortcut backend after resume to keep shortcuts working (PR #371)
Delayed keyboard restart to allow USB re-enumeration after resume (PR #372)
Fixed premature transcription during push-to-talk silence (fixes #358, PR #359)
Disabled copy-to-clipboard by default in Settings (PR #370)
Maintenance updates: npm/yarn dependency refresh and brace-expansion dev dependency bump (PR #346, #357)

View release on GitHub

v0.10.0-beta

Beta2026-03-25

Generalized keyboard modifier alias matching across layouts for more reliable shortcuts
Audio channel probing now validates device-supported sample rates before selection
evdev now handles SYN_DROPPED to prevent stale modifier state
IBus engine activation now uses register_component for stronger text-injection startup
Settings dialog forces window decorations to prevent missing-titlebar behavior
Tray icon refresh now uses icon names for better AppIndicator compatibility
Coverage increased to 80%+ with additional IBus launch/main-entry tests
Installer and CI polish: latest-tag fallback via GitHub API, Node 24 deploy, and path-filtered workflows

View release on GitHub

v0.9.0-beta

Beta2026-03-14

Left/right modifier key distinction - choose Left Ctrl vs Right Ctrl for your shortcut
Sound effects toggle - enable or disable audio feedback from Settings
Wayland clipboard fallback - auto-copies text when virtual keyboard injection isn't available
Display availability check - graceful error when running in headless environments
Fixed unwanted leading space at the start of each new transcription session
Fixed shortcut mode (toggle/push-to-talk) not applying on startup
Improved Debian/pipx installation guidance and cross-distro error messages
Grouped shortcut selector UI - shortcuts organised by Either/Left/Right side

View release on GitHub

v0.8.0-beta

Beta2026-03-01

Push-to-talk shortcut mode (hold to speak, release to stop)
Optional voice commands with VOSK auto-enable behavior
Improved shortcut mode switching and callback reliability
IBus active-method detection before text injection
Audio hardware compatibility fixes (sample rate and channel count)
Fedora startup dialog stability fix
Web SEO expansion and homepage visual refresh

View release on GitHub

v0.7.0-beta

Beta2026-02-22

Autostart on login support (XDG autostart)
Tabbed settings dialog (Speech Engine, Recognition, Text Injection, Audio Feedback, General)
Intel GPU compatibility detection - auto fallback to CPU for incompatible GPUs
Single instance prevention - prevents multiple Vocalinux running
Evdev device management - removes disconnected devices to prevent CPU spin
Improved GPU detection - avoids false positives on systems without dev libraries
IBus fallback - skip setup when daemon not running
Fedora dnf check-update fix
Web SEO - 7 new optimized pages