d11bd1bf0d
* feat(post-processing): pipe transcriptions through user-defined script Add a configurable post-processing step that passes each transcription result to an external executable via stdin and uses its stdout as the text to type. Enables custom transforms (formatting, LLM rewriting, etc.) without modifying Vocalinux. Falls back to original text on script failure or timeout. * test(post-processor): add unit tests and extract apply_post_processing Move config-reading logic from main.py closure into apply_post_processing() in post_processor.py so it can be tested independently. Add 11 unit tests covering PostProcessor.process() and apply_post_processing(). * docs(post-processing): document post-processing script feature * fix formatting of one file * fix(post-processing): run the user script with a stripped host environment Rebase compat: the host_process AST guard requires every subprocess spawned for a host binary to pass env=host_env() so an AppImage's LD_LIBRARY_PATH/GI_TYPELIB_PATH/PYTHONHOME do not leak in. Also give apply_post_processing() a typed ConfigManager parameter and keep text_callback_wrapper's local variable typed as str for mypy. * fix(post-processing): preserve paragraph breaks and tighten script handling Take script stdout verbatim instead of rstripping every trailing newline, so the "new paragraph" voice command's break survives a passthrough script. A single trailing newline is still dropped, but only when the transcription had none — it is almost always just the line terminator echo/print-style tools append. Also stop writing transformed output to the log (it can carry sensitive dictation into retained logs), catch only the expected process/OS/encoding failures instead of every Exception, and add the missing signature annotations. * fix(main): run the post-processing script off the recognition thread A user script may take seconds per segment, and text callbacks run on the recognition thread that drains the bounded audio-segment queue; stalling it drops queued dictation once the queue fills. Submit post-processing and injection to a dedicated single-worker executor so the backlog waits instead of losing speech, while keeping injection order. On timeout or failure the worker falls back to the original text, and an unexpected error is logged with a traceback and still injects the original rather than silently dropping the segment. Also pin the callback tests' get_str mock to no-script so they exercise the synchronous path deterministically, and add a callback test proving transformed text (including preserved paragraph breaks) reaches the injector. * docs(post-processing): annotate new callbacks and clarify stdout contract Add the missing type annotations on the new Settings page callbacks per the repository guide, and document that script stdout is used verbatim (a lone trailing newline added by line-oriented tools is dropped only when the transcription had none) and that scripts run on a dedicated worker. * test(main): annotate new post-processing callback tests * fix(main): bind queued dictation to its target app and drain worker on quit Route every transcription segment through the single post-processing worker (a pass-through when no script is configured) so clearing the script path mid-queue can no longer let a later segment overtake an earlier one still running. A segment that can sit in the backlog — a script is configured or a job is still running — is bound to the application focused when it was dictated; the worker drops it if focus has verifiably moved to another application, so delayed dictation cannot land in the wrong app. Wire the worker into the tray quit path via an on_quit hook: it clears the acceptance flag, cancels pending jobs, and waits — bounded by the script's own timeout — so nothing injects after the injector stops. * fix(main): keep voice actions ordered and probe focus off the recognition thread - Queue voice-command actions on the post-processing worker so an action can no longer overtake dictated text still waiting to be injected. - Move the submit-time focus probe onto a dedicated executor; the recognition thread only submits the probe, and the job joins the result at inject time, so compositor tools can no longer stall consumption of the bounded audio-segment queue. - Re-check the quit flag after the inject-time focus re-probe, then shut both workers down without waiting on a running script so quitting does not freeze on the script timeout. * fix(main): capture submit-time focus on dedicated threads and serialize quit - Replace the shared focus-probe executor with a fresh daemon thread per submission so a slow earlier probe cannot delay a later one past the focus change it is meant to capture; jobs read the result off a one-slot queue at inject time. - Give queued voice actions the same submit-time focus binding as text segments so a queued shortcut is dropped instead of fired into a newly focused application. - Hold injection_lock across the accept-check plus inject/action call so quit waits only for an in-flight injection instead of racing it. - Queue the IDLE reset of last_injected_text behind pending worker jobs so a queued 'delete that' still sees the text it targets. * fix(main): keep actions instant on a free worker and bound quit wait - An action now starts a submit-time focus probe only when it actually queues behind other jobs; with a free worker there is no delay window to bind against, so the probe's compositor calls are skipped and the action stays immediate even when a script is configured. - Probe and executor submission now happen inside pending_jobs_lock so queue order matches callback order — a job that saw an empty queue can no longer slip behind one that arrived while it was submitting. - Every inject-path subprocess call (xdotool window/key/type, wtype and ydotool type) is bounded by a timeout, and the quit hook's injection_lock wait is capped, so quitting can no longer freeze on a stalled injector subprocess. * fix(main): bind every action to submit-time focus without slowing it down - Every voice action now starts a submit-time focus probe: submission never guarantees immediate execution, so skipping the binding left a focus-change window in which a shortcut could fire into another app. The worker waits on the probe for at most one second — milliseconds on a healthy desktop — so immediate actions stay immediate while a genuinely stalled probe falls back to the injector's own targeting. - Quit waits for an in-flight injection without a cap again: every inject-path subprocess is individually bounded, so the wait always ends with the injection and a long transcription is never cut off. - The wtype/ydotool typing timeout is now derived from the configured per-key delay (a generous multiple), so a legitimately slow type can never be cut mid-text; a timed-out type no longer falls through to the clipboard fallback, which would duplicate the partial fragment. * fix(main): drop unverified actions and make quit-time injection abortable - An action whose submit-time focus probe cannot answer inside its one-second bound is now dropped instead of running unverified: a stalled probe can no longer leave a shortcut pointing at whatever application happens to be focused. - Quit asks the injector to abort in-flight injections, and injection paths honour it: the xdotool chunk loop stops between chunks, the wtype/ydotool type call runs as a polling Popen that can be killed, and paste/shortcut/backspace entry points exit early. The quit hook still waits for an in-flight injection, but the wait is bounded by at most one already-running subprocess call — never the injection's full length — and no subprocess is ever stopped mid-call by a timeout cap on the wait itself. - Annotate the remaining untyped test helpers. * fix(main): wait out slow focus probes; restore run() in wayland typing - An action no longer races its submit-time focus probe on a one-second bound: the job waits for the probe's own answer, whose compositor calls already carry per-call timeouts and short-circuit absent tools. A slow but living probe delays the command instead of discarding it, so a valid "undo" can no longer be lost to timing alone, and focus verification still decides every drop. - Wayland typing goes back to subprocess.run so tests keep intercepting the same entry point. Abortability is preserved by typing in 200-character chunks with the shutdown check between calls — a whole-string call could not be interrupted, but no single chunk holds the injector for more than its own scaled budget. - The abort flag gains a class-level default so injectors built without __init__ in tests still answer the checks; each real instance gets its own event in __init__. - Update call-shape assertions for the typed-call timeout and finish annotating the test methods touched by the callback refactor. * fix(inject): never retype text a failed chunked call already sent - _PartiallyTyped carries the confirmed-delivered prefix length out of the chunked type loops, so every downstream fallback continues from the untyped remainder: the Wayland-to-xdotool switch types only the tail, and the clipboard fallback copies only the tail, never the whole string a second time. - The xdotool retry loop gains the same accounting — a retry now resumes at the first untyped chunk instead of restarting at the top and duplicating everything the failed attempt already delivered. - Annotate the remaining test signature the repository guide requires. * fix(inject): report partial typing as failure even when clipboard lands A chunked type call that dies mid-text leaves only a prefix on screen, and the remainder-fallback keeps it that way: it copies the tail for a manual paste but now returns False, so the caller no longer records the full transcription as injected and a following "delete that" cannot erase text that was never typed. * fix(inject): keep deletion state aligned with delivered text inject_text now records how many characters a call confirmed on screen as last_typed_count: the whole length on success, the typed prefix after a partial failure, -1 when the delivered amount is unknowable. The transcription caller uses it to keep last_injected_text honest — a following "delete that" erases the typed prefix or nothing, never the previous segment or text that never arrived. * fix(inject): never replay input a timed-out portal job may have delivered - RemoteDesktopPortalError.delivered=None marks an unreported job's count as unknowable; all three portal fallbacks demote and skip instead of replaying, and text injection surfaces as TimeoutExpired so no full-text clipboard copy is offered - _PartiallyTyped counts raised after a portal demote are rebased onto the original text (typed_prefix) so the caller clips the real prefix - evdev: a consumed pure-modifier press is withheld and replayed ahead of the first non-target press, so holding RightAlt for AltGr composition still reaches the application - history: extend_latest gains an expected_latest CAS guard so a late segment from an older worker merges only into its own session's snippet, never into a newer session's entry - quit path aborts in-flight injections before waiting on the lock so quitting cannot freeze behind a long transcription's typing budget - tests: stop _ensure_ydotoold leaking a real Popen in Wayland tests * fix(shortcut): replay withheld modifiers across devices and after SYN_DROPPED - Replay withheld modifier presses from every device, not only the fd the character arrived on: split keyboards deliver AltGr and the character on separate devices, and a per-fd check forwarded a plain character. - Let the post-SYN_DROPPED resync run when only withheld modifiers exist for the fd, so a press whose release the dropped burst ate is pruned instead of replaying a stale modifier on the next ordinary key. * fix(history,pads): keep late segments pad-bound deliveries and pad deletions correct - Track the worker behind the ended session's snippet so late segments from an older worker cannot merge into a newer session's entry - Skip the focus-change drop for dictation-pad segments and pad-bound editing commands, which never target the focused application - Refuse 'delete that' when a manual pad edit blurred the segment boundaries instead of deleting the stale recorded length * style: drop unneeded nonlocal for mutated worker set * fix(shortcut): share one evdev device layer across shortcut listeners Each KeyboardShortcutManager built its own EvdevKeyboardBackend, and on evdev every backend opened and grabbed the same physical keyboards: the first grabber received all events and the rest received none, so per-language shortcuts could never fire on Wayland. All backends now bind to a process-wide EvdevDeviceHub that opens each keyboard once, owns the grabs and the paired uinput clones, and fans every event out to each registered engine. An event is re-emitted to applications only when no engine consumed it, so suppression semantics stay the same while per-engine matching state (combo progress, withheld modifiers, swallowed presses) remains independent. * fix(history): own late segments per worker and decide pad routing once - ended_worker_snippets maps each ended session's worker to its own (snippet, commit epoch): late deliveries merge into their session's entry however many sessions have committed since, so one session's stragglers never split across Recent Snippets entries. - The leftover-segment commit reads the live history epoch inside the session lock: a mid-session clear() no longer makes the stale ended epoch refuse valid post-clear dictation. - The pad-bound routing decision is captured once per job and passed to the injector/action handler, so a toggle flip between the focus check and delivery cannot reroute text or shortcuts into the focused app. - Dead recognition workers are pruned from the tracking set and map so long runs don't accumulate finished threads. - The custom-terms switch now suppresses GTK's own state-set on save failure so it shows the value that actually persisted. * fix(shortcut): serialize evdev device layer teardown against restart The last listener's unregister released the engine lock before joining the monitor thread and closing devices, so a new listener registering in that gap opened keyboards the stale teardown then closed — it stayed active but dead, and briefly grabbed keyboards stopped forwarding input. A lifecycle lock now covers the whole close/open transition in register, unregister, and the monitor's unexpected-exit teardown, so a cold start can only open devices after the previous generation finished closing (also making the new generation's grabs succeed). A generation stamp tells a monitor thread that outlived its join() its incarnation is over even when running has gone True again. * fix(injection): log focus-probe failures and narrow clipboard fallback catch A crashed focus probe turned into an unknown result with no logged error, indistinguishable from unavailable focus information; the partial-typing clipboard fallback caught every exception. Log the probe failure with context and narrow the clipboard catch to the exceptions the fallback path can actually raise. * fix(shortcut): join the evdev monitor outside the lifecycle lock If the monitor exited unexpectedly while the last listener unregistered, its cleanup queued on the lifecycle lock that unregister() held for the join — stalling shutdown for the join's full 2s timeout. The close now finishes inside the lock and the join waits after it is released, so the monitor's own cleanup can run and exit instead of queueing behind it. Closing before the join also dies the monitor's read/select on the closed fds right away. A _closed gate on _open_keyboard_device keeps a rescan from opening a device between detach and close that nothing would ever close. * fix(injection): narrow focus-probe catch to the raisable exception set get_focused_window() reports Optional and never raises; the catch now covers only what can realistically escape the queue write and logs it with context, keeping the waiting job released. * fix(shortcut): tag evdev device fds with the generation that opened them A monitor that outlived join() could still touch a newer generation's devices: the loop re-checked the generation only at the top, so mid-iteration it could dispatch from, forward to, or remove devices a restarted listener had just opened. Every fd is now tagged with the generation that opened it, and the tag is verified at each point a monitor could mutate shared state — fd lookup for dispatch, clone forwarding, forwarder release, device removal, and rescan opens — while dispatch re-checks the generation between events and the per-fd loop breaks on it. Cold starts bump the generation before opening so fresh devices are tagged for the new generation. Devices an older generation opened stay tagged stale, so its reader leaves them alone. --------- Co-authored-by: Jatin K Malik <jatinkrmalik@gmail.com>
Vocalinux turns speech into typed text in whatever app has focus. It is a free, AGPL-3.0-licensed desktop app for X11 and Wayland. After you download a model, local engines (whisper.cpp by default, plus OpenAI Whisper, Faster Whisper, VOSK, and Parakeet) run speech-to-text on your machine. An optional remote HTTP API is off unless you configure it.
No Voca account is required. Models download once. After that, speech-to-text stays on your machine.
Current release: v0.17.0. Minor: Faster Whisper and Parakeet engines, Speech Model simple setup, first-run system language, Snap with native Wayland typing, Flatpak bundles on the GitHub Release. Details: docs/UPDATE.md.
Features
- On-device after model download: Local engines; speech-to-text stays on your machine
- X11 and Wayland: Text injection via xdotool, IBus, wtype, ydotool, or clipboard fallback
- Several engines: whisper.cpp (default), OpenAI Whisper, Faster Whisper, VOSK, Parakeet, plus optional remote HTTP API or local VocaGateway from Settings (not on-device)
- GPU acceleration: Vulkan for AMD, Intel, and NVIDIA with whisper.cpp
- Toggle or push-to-talk: New installs default to hold Right Alt; existing configs keep their shortcut
- System tray + settings: Searchable sidebar, Speech Model simple setup with Advanced as an island, status icons, audio feedback
- Start on login: XDG autostart (desktop session, not a systemd service)
- Packaging: install script, AppImage, AUR, PyPI, Snap (
--edge), Flatpak (release bundles and local build; not on Flathub)
Screenshots
Vocalinux in action. Full gallery on the website screenshots page.
Product
 Real-time voice-to-text transcription |
 System tray with listening indicator |
 About & Updates in Settings |
 Log viewer for debugging |
Settings
 Speech Engine |
 Recognition |
 Audio |
 Performance |
 General |
 Advanced |
 Floating overlay while listening (green) |
 Floating overlay while processing (amber) |
Install
Recommended (interactive installer)
curl -fsSL https://raw.githubusercontent.com/VocaHQ/vocalinux/main/install.sh -o /tmp/vl.sh
bash /tmp/vl.sh
Prefer to review the script first: open /tmp/vl.sh before running it, or clone the repo and run ./install.sh locally.
The installer detects hardware, recommends an engine, downloads a default model (~74MB for whisper.cpp tiny), installs neural VAD when ONNX Runtime is available, and sets up desktop integration. Typical install time with whisper.cpp is about 1-2 minutes.
| Engine | When to use |
|---|---|
| whisper.cpp (default) | Best default; Vulkan GPU on AMD, Intel, and NVIDIA |
| Whisper (OpenAI) | PyTorch path; NVIDIA/CUDA |
| Faster Whisper | CPU-friendly Whisper via CTranslate2 / INT8 |
| VOSK | Low RAM / minimal footprint |
| Parakeet | CPU; NVIDIA NeMo ASR via sherpa-onnx; 25 European languages |
| Remote API | Offload to a server you configure |
Non-interactive options:
bash /tmp/vl.sh --auto # whisper.cpp defaults
bash /tmp/vl.sh --auto --engine=whisper # OpenAI Whisper
bash /tmp/vl.sh --auto --engine=faster_whisper # Faster Whisper (CPU)
bash /tmp/vl.sh --auto --engine=vosk # VOSK only
bash /tmp/vl.sh --auto --engine=parakeet # Parakeet (CPU)
For a specific release tag, see GitHub Releases or ./install.sh --tag=v0.17.0.
Arch Linux (AUR)
yay -S vocalinux
See docs/AUR.md.
AppImage
Download the x86_64 or aarch64 AppImage from Releases, mark it executable, and run it. Built against glibc 2.35 (Debian 12+, Ubuntu 22.04+, Fedora 36+, Arch, Tumbleweed). Host text-injection tools (xdotool on X11; wtype / ydotool / clipboard tools on Wayland) are still required. Current AppImages rebuild whisper.cpp with Vulkan and use the host GPU driver. Prefer the installer when you want system deps, a CUDA build, and models set up automatically.
Flatpak (any distro)
Install and auto-update via the self-hosted VocaHQ remote (flatpak update picks up each release):
flatpak install https://vocahq.github.io/vocalinux-flatpak/com.vocalinux.Vocalinux.flatpakref
The .flatpakref adds the vocahq remote and resolves the GNOME runtime through Flathub. Manually instead:
flatpak remote-add --if-not-exists flathub https://dl.flathub.org/repo/flathub.flatpakrepo
flatpak remote-add --if-not-exists vocahq https://vocahq.github.io/vocalinux-flatpak/vocahq.flatpakrepo
flatpak install vocahq com.vocalinux.Vocalinux
Or sideload a GitHub Release bundle (Vocalinux-<version>-x86_64.flatpak or -aarch64.flatpak) after the Flathub GNOME runtime is present — bundles do not auto-update:
flatpak install --user ./Vocalinux-<version>-x86_64.flatpak
Local build:
flatpak install flathub org.gnome.Platform//50 org.gnome.Sdk//50
flatpak-builder --user --install --force-clean build-dir \
packaging/flatpak/com.vocalinux.Vocalinux.yml
flatpak run com.vocalinux.Vocalinux
Ships whisper.cpp with Vulkan. It is not on Flathub (submission flathub#9368 closed 2026-07-23 on policy grounds). Details: packaging/flatpak/README.md.
Snap (Ubuntu Snap Store)
Listing: snapcraft.io/vocalinux. stable is promoted from candidate after QA. Tagged releases also attach vocalinux_<version>_amd64.snap on GitHub.
sudo snap install vocalinux --edge
sudo snap connect vocalinux:audio-record # if mic is not auto-connected
sudo snap connect vocalinux:raw-input # global keyboard shortcuts (evdev)
sudo snap connect vocalinux:hardware-observe # list keyboards (/proc/bus/input/devices)
sudo snap connect vocalinux:uinput # native Wayland typing (ydotool)
v0.17.0 includes ydotool and the uinput plug. The Store still has to finish review of that plug, so snap info vocalinux may still show v0.16.2 (rev 7) on edge. That revision has no uinput plug; snap connect vocalinux:uinput fails until a 0.17.0+ revision is listed.
Until then, sideload the GitHub file:
sudo snap install --dangerous ./vocalinux_0.17.0_amd64.snap
sudo snap connect vocalinux:uinput
From source
git clone https://github.com/VocaHQ/vocalinux.git
cd vocalinux
./install.sh
# or pick the engine up front
./install.sh --engine=whisper_cpp
./install.sh --engine=vosk
./install.sh --engine=parakeet --auto
After installation
vocalinux # if ~/.local/bin is on PATH
# or
~/.local/share/vocalinux/venv/bin/vocalinux
You can also launch Vocalinux from your application menu.
Nightly builds
Daily builds from main appear on Releases. Use the latest stable or beta release for production; nightlies are untested.
Requirements
| OS | Linux (Ubuntu 24.04+, Debian 12+, Fedora 42+, Arch, openSUSE Tumbleweed) |
| Python | 3.11 or newer |
| Display | X11 or Wayland |
| Hardware | Microphone; GPU optional (Vulkan) |
The distro must ship Python 3.11+ because Vocalinux uses distro PyGObject (python3-gi). Ubuntu 22.04 (Python 3.10) and Debian 11 (3.9) are below that floor. Distribution notes: docs/DISTRO_COMPATIBILITY.md.
Usage
Dictation
- Push-to-talk (default on new installs): hold Right Alt (Option on Mac-layout keyboards) and speak
- Speak into the microphone
- Release to stop, or switch to Toggle mode in Settings (double-tap a key to start/stop)
Existing configs keep their saved shortcut.
Voice commands
English phrases always work. With a non-English recognition language, matching punctuation / line-break phrases in that language are also recognized (Italian virgola / punto, French virgule / point, etc.).
| Command | Action |
|---|---|
| "new line" | Line break |
| "period" / "full stop" / "dot" | . |
| "comma" | , |
| "question mark" | ? |
| "exclamation mark" | ! |
| "delete that" | Delete last sentence |
| "capitalize" | Capitalize next word |
CLI
vocalinux --help
vocalinux --version
vocalinux --debug
vocalinux --engine whisper_cpp # default
vocalinux --engine whisper
vocalinux --engine faster_whisper
vocalinux --engine vosk
vocalinux --engine parakeet
vocalinux --engine remote_api
vocalinux --model medium
vocalinux --model medium.en-q5_0 # exact whisper.cpp variant
vocalinux --model large-v3-turbo
vocalinux --wayland
vocalinux --start-minimized
Autostart
Start on Login creates an XDG autostart desktop entry (~/.config/autostart/). It does not install a systemd unit. Enable from the first-run dialog, tray menu, or Settings.
Configuration
Stored at ~/.config/vocalinux/config.json. Prefer the Settings dialog for day-to-day changes. Settings → Speech Model starts with a simple setup; expand Advanced for engine, size, and specialization. Set post_processing.script_path to an executable to transform each transcription result (stdin → stdout) before it is injected; leave empty to disable.
Neural VAD (Silero) is used when onnxruntime is available; install via pip install "vocalinux[vad]" for manual/PyPI installs. The installer attempts this automatically.
Documentation
| Document | Description |
|---|---|
| Installation | Installer, AppImage, AUR, Flatpak, Snap, running |
| Manual / PyPI install | Package lists and pip workflows |
| Troubleshooting | Common failures |
| User guide | Dictation, engines, models, tips |
| Update guide | Upgrade steps and release notes |
| Changelog | Release history pointers |
| Support | Where to get help |
| Distribution compatibility | Distro matrix and session notes |
| Remote HTTP API | Offload transcription to a server |
| Local VocaGateway (optional) | Start a local gateway from Settings; not on-device |
| Contributing | Dev setup, style, PR process |
| Security | Supported versions and vulnerability reporting |
| Docs index | Full documentation map |
Privacy and security
- Local engines process audio on-device after you download a model; no Voca account required
- Optional remote API is off by default and only used when you configure a server
- Model downloads are checked against pinned checksums
Report vulnerabilities privately per SECURITY.md.
Development
git clone https://github.com/VocaHQ/vocalinux.git
cd vocalinux
./install.sh --dev
source venv/bin/activate
pytest
python -m vocalinux.main --debug
See CONTRIBUTING.md for the two-venv layout, just recipes, and PR guidelines.
Roadmap
Shipped: graphical settings, multi-language support, whisper.cpp default, Vulkan GPU, Wayland/IBus, Flatpak packaging, AppImage, in-app update checker, Parakeet and Faster Whisper engines, Snap recipe.
Planned:
- Application-specific voice commands
- Debian/Ubuntu package (
.deb) - User-customizable voice command map
- Flathub publication (not currently listed; see #167)
Voca ecosystem
Vocalinux is part of VocaHQ. On-device speech-to-text first, one app per platform. Optional VocaGateway is self-hosted and not on-device.
| Platform | Project | Website | GitHub |
|---|---|---|---|
| Linux | VocaLinux | vocalinux.com | VocaHQ/vocalinux |
| macOS | VocaMac | vocamac.com | VocaHQ/vocamac |
| Windows | VocaWin | vocawin.com | VocaHQ/vocawin |
| Phone | VocaPhone | vocaphone.vocahq.com | VocaHQ/vocaphone |
| Gateway | VocaGateway | vocagateway.vocahq.com | VocaHQ/vocagateway |
Talk to us: Discord · X @vocahq · hello@vocahq.com
Contributing
Bug reports, docs, and code are welcome. Start with CONTRIBUTING.md and good first issues.
Contributors
Thanks to everyone who has contributed code, docs, or fixes:
Repository mirrors
GitHub is the primary forge for issues, PRs, CI, and releases.
| Role | URL |
|---|---|
| Primary | https://github.com/VocaHQ/vocalinux |
| Read-only mirror (Codeberg) | https://codeberg.org/jatinkrmalik/vocalinux |
Open issues and pull requests on GitHub only.
License
This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0), aligning with the other VocaHQ distribution projects.
You may use, study, modify, and redistribute the software under AGPL-3.0. If you run a modified version as a network service, AGPL also requires that you make the corresponding source available.
Star Chart
Optional Orukeet model
With the Parakeet engine installed, Settings → Speech Model → Parakeet also
offers orukeet-v0.1.0. It uses the existing local sherpa-onnx recognizer for
25 European languages; the recommended Parakeet model and default whisper.cpp
engine stay unchanged.
The download comes from a pinned revision of oruk/orukeet on Hugging Face. Vocalinux checks every file against its bundled SHA-256 pins and checks the publisher's release manifest against those same pins. The manifest download participates in Hugging Face's normal model-download accounting. Cached recognition works offline and does not send audio or usage events. The model occupies about 641 MiB; its CC BY-SA 4.0 weights license and NVIDIA attribution are included.