SnipVoice v1.2.1

SnipVoice v1.2.1 adds a Voice Commands editor and handsfree shortcuts, and improves recording shutdown and recovery under load.

What's changed

• Dictation > Voice Commands: manage spoken phrases and the literal text each command inserts.
• Hold to talk or handsfree Start / Stop: choose how the dictation and voice-command shortcuts work. In handsfree mode, press once to start and again to stop.
• Reliable shortcut ordering: rapid press, release, and Escape events stay ordered. Excessive shortcut input cancels the gesture with a retry message, and cancelled takes cannot start obsolete inference.
• Audio shutdown: accepted dictation audio drains after a queue stall, delayed pause commands preserve the stopping state, and completed worker records no longer accumulate.
• Meeting recordings: contiguous Windows audio packets are batched to reduce disk synchronization pressure, and queued audio is preserved when stopping races the native helper's exit. Overflow and unexpected exits remain visible.
• Floating recording indicator: recording time, pause state, and partial/interrupted status stay visible while the manager is minimized or hidden.
• Model selectors: descriptive names are ordered from lower to higher quality in Dictation, Recording settings, and Models.
• Summary citations: repeated valid transcript references are normalized; unknown references are still rejected.

Downloads

• Windows x64: per-user installer and portable ZIP.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and platform build-information archives accompany the clean LGPL audio runtime.
• SHA-256 manifests are included for every download.

Verification

• Full local Windows suite: 1,231 tests ran successfully, with 10 platform/environment skips.
• Post-merge CI passed on Windows, macOS, and Linux. The delayed-library GUI fixture race found during release checks was reproduced and fixed.
• Tagged desktop builds passed on Windows x64 and macOS ARM64, including the clean MP3 codec checks, packaged native runtime probes, and macOS signature verification.
• Downloaded asset checksums, archive layouts, and macOS version/channel metadata were verified. The downloaded Windows package passed voice, summary, capture-helper, and SQLite probes plus cold startup with disposable data and profile directories.

The packages retain the existing signing status: the Windows installer is not publisher-signed, and the macOS bundle is not notarized. Physical microphone/loopback capture and installed-app operation were not re-tested for this release.

See the full changelog.

SnipVoice v1.2.0

SnipVoice v1.2.0 makes a recording's audio files explicit, builds the final audio about 10x faster, fixes the microphone boost, and lowers SnipVoice's memory and CPU use while idle.

What's changed

• Library > Files separates the final audio from the original (raw) tracks and shows the format and size of each.
• Download final audio saves a copy of the audio used for playback, in its original format. It no longer needs the original tracks.
• Download each original track (microphone or system audio) on its own, as MP3.
• Remove original tracks frees most of a recording's space: they are uncompressed, often around 30 MB per minute. It removes both tracks together and is only offered once a final audio file is saved. The confirmation warns that the recording can no longer be transcribed again. Afterwards, playback and downloads use only the final audio, and the transcript, summaries, chat, and citations stay. Automatic retention in Settings > Privacy follows the same rule.
• Original tracks are no longer pre-selected for removal, and removal never falls back to every track when none is chosen.
• Automatic clean-up: Settings > Privacy can remove original tracks after a number of days you choose (30 by default when turned on). SnipVoice applies it at startup and every six hours, only to recordings with a saved final audio and a finished transcript. The previous raw-audio policy was saved but never applied.
• Settings > Privacy in plain language: every option is a sentence with a short hint instead of internal codes. "Automatically move recordings to the trash after N days" now actually runs in the same pass, and moves recordings to the trash, where they can still be restored. Nothing is purged automatically.

Faster, better final audio

• About 10x faster: after a recording, the final audio is ready in about 20 seconds instead of 3-5 minutes for a 24-minute meeting. Adjust volume and audio downloads get the same speed-up.
• Mic boost that holds up: one desk knock or laugh no longer switches the boost off for the whole recording. Loud moments where you and the call overlap are limited instead of clipped.
• Balanced against the call: when the call is louder than you, your voice is raised to at least the call's speech level (up to 8x).
• Adjust volume rebuilds the MP3 in place instead of leaving "(2).mp3" copies.

Lighter in the background

• One model by default: recordings are transcribed with your dictation model unless you choose another in Settings > Recording ("Same as dictation").
• Memory: the dictation model (about 0.8 GB for the default) is released after 10 minutes without use. Quick follow-ups stay instant; the first dictation after a quiet spell waits about a second while it reloads.
• CPU: idle use while the window is closed dropped from about 1.3% to 0.3% of a core.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.

SnipVoice v1.1.0

SnipVoice v1.1.0 redesigns the manager window so pages use its full height instead of scrolling.

What's changed

• Sidebar navigation. The app icon and name, the Recording, Library, Dictation, and Settings pages, and the 100% local badge now sit in a left sidebar. The header, tab strip, and per-page titles are gone. Ctrl+1 to Ctrl+4 still switch pages.
• Two-column Recording page. The title and sources are on the left. The live waveform, level meters, a large clock, and Start/Pause/Stop are on the right, so you check the signal and start in one place.
• Library. Search and the library actions sit above the recording list, so the selected recording gets the pane's full height. The player sits above the transcript and replaces the separate Listen tab.
• Fewer scrollbars. Pages show a scrollbar only when their content overflows, and long device names fit in the macOS source pickers.
• The minimum window width is now 1040 px on Windows and 1100 px on macOS to make room for the sidebar.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.

SnipVoice v1.0.0

SnipVoice v1.0.0 is the first stable release: local push-to-talk dictation and a meeting recorder with separate microphone and system-audio tracks, now with an English interface.

What's changed

• The interface is available in Brazilian Portuguese (default) and US English. Switch in Configurações > Geral > Idioma (Settings > General > Language); the window rebuilds without restarting.
• The manager follows the Win Design System in light and dark appearance, with Ctrl+1 to Ctrl+4 to jump between Recording, Library, Dictation, and Settings.
• A three-second source check with live meters runs before recording, and saved recordings replay with source selection and seeking.
• Quiet microphone speech is raised in the final recording without touching the recoverable raw tracks.
• The local summary runtime runs isolated from transcription's native audio libraries.

Versioning

Earlier releases were renumbered for the stable launch: 1.0.0–3.3.2 are now 0.1.0–0.6.2, and the 3.4.0 betas are the 1.0.0 betas. They point at the same commits.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.

SnipVoice v1.0.0 beta 3

Formerly v3.4.0-beta.3; SnipVoice releases before 1.0.0 were renumbered.

Snipvoice v3.4.0 beta 3

This beta brings the manager and recording library into the Win Design System. Recording controls and level meters remain reachable at the minimum window size, library playback stays prominent, and less-used dictation tools move into a secondary menu.

The beta also includes the recording and runtime improvements from beta 2: quieter microphone speech is raised in saved audio, and local summaries run separately from transcription's native libraries.

Downloads

• Windows x64: installer or portable ZIP
• macOS ARM64: ad-hoc-signed app ZIP
• SHA-256 checksums and FFmpeg source/build information accompany the bundles

There is no MSIX in this beta. Beta 2 already used Store package version 3.4.0.0, and another package at that version cannot be submitted as an update.

These packages are not publisher-signed or notarized. Real-device microphone and system-audio recording feedback is still needed before the stable release.

Full changes: v3.4.0-beta.2...v3.4.0-beta.3

SnipVoice v1.0.0 beta 2

Formerly v3.4.0-beta.2; SnipVoice releases before 1.0.0 were renumbered.

SnipVoice v3.4.0 beta 2 fixes the Windows local-summary startup error, improves quiet microphone recordings, and adds replay to the meeting library. The current stable release remains v3.3.2.

What's changed

• Run local summaries in a separate process so the native llama.cpp library does not collide with transcription's audio libraries on Windows.
• Raise quiet microphone speech in the derived final recording while preserving the original raw track.
• Replay saved recordings from the library, choose an available source, and seek through the recording. The library opens with the useful details visible and advanced tools tucked away.
• Display the product name as SnipVoice in the app and installer.
• Add a Microsoft Store MSIX build path for Partner Center submission. Its unsigned package is a submission artifact, not a direct-install download.

Downloads

• Windows x64: portable ZIP or per-user installer.
• macOS 14.4+ ARM64: app bundle ZIP.
• Microsoft Store: unsigned MSIX for Partner Center submission only; use the Windows installer or ZIP for direct installation.
• SHA-256 files accompany the packages. FFmpeg source and build-information archives accompany the clean LGPL audio runtime.

This beta uses the same installation and data location as the stable build. Back up your data before testing it. The packages are not notarized or publisher-signed.

Automated source tests and bundle builds passed. The Windows portable package passed native runtime probes. Physical microphone and system capture, packaged playback, real-model summary generation, and the installed GUI still need real-device verification before a stable release.

See the full changelog.

SnipVoice v1.0.0 beta 1

Formerly v3.4.0-beta.1; SnipVoice releases before 1.0.0 were renumbered.

Snipvoice v3.4.0 beta 1 previews the new recording workflow. The current stable release remains v3.3.2.

What's changed

• Test microphone and system sources for three seconds before recording, with live levels and a waveform. The check does not save audio or create a meeting.
• Use a simpler recording screen with title, source selection, waveform, and transport controls. Destination and automation defaults are under Configurações > Gravação.
• Double-click a transcript row or press Enter to play from that point, including while another segment is playing.

Downloads

• Windows x64: portable ZIP or per-user installer.
• macOS 14.4+ ARM64: app bundle ZIP.
• SHA-256 files accompany the packages. FFmpeg source and build-information archives accompany the clean LGPL audio runtime.

This beta uses the same app installation and data location as the stable build. Back up your data before testing it. The packages are not notarized or publisher-signed.

The automated tests and package probes passed, but physical microphone/system capture and packaged playback have not yet been verified on real devices. Please report those results before a stable release.

See the full changelog.

SnipVoice v0.6.2

Formerly v3.3.2; SnipVoice releases before 1.0.0 were renumbered.

Snipvoice v3.3.2 adds more local transcription models, movable data and model folders, and a clearer model settings page.

What's changed

• Whisper Small, Whisper Large v3 Turbo and Whisper Large v3 are available as optional, hash-pinned local transcription models.
• Configurações > Geral > Pasta de dados moves all Snipvoice data to another folder; the app restarts and verifies the copy before removing the old folder.
• Configurações > Modelos > Pasta dos modelos moves downloaded models to another folder, including one shared with other apps.
• Transcription and summary model downloads are grouped under Configurações > Modelos, and download buttons stand out clearly on their cards.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.

SnipVoice v0.6.1

Formerly v3.3.1; SnipVoice releases before 1.0.0 were renumbered.

Snipvoice v3.3.1 is a Windows polish release.

What's changed

• The Windows icon now fills tray, Start and taskbar slots instead of looking smaller than neighbouring icons.
• The Windows window pump keeps processing queued work even when Tk timer wake-ups stop arriving.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.

SnipVoice v0.6.0

Formerly v3.3.0; SnipVoice releases before 1.0.0 were renumbered.

Snipvoice v3.3.0 makes the library and settings easier to navigate and fixes appearance defects across both themes.

What's changed

• The library is easier to navigate: one toolbar, filters and cross-meeting questions on demand, and each recording split into Notas, Transcrição, Resumo, Perguntar and Arquivos.
• Configurações is organized into Geral, Privacidade, Transcrição and Resumos sections instead of one long page.
• Dark and light themes render checkboxes, fields, meters, scrollbars and borders correctly, and controls stay whole at the minimum window size.
• The clean Windows audio runtime can be built from PowerShell with an MSYS2 toolchain.

Downloads

• Windows x64: portable ZIP and per-user installer.
• macOS 14.4+ ARM64: ad-hoc-signed app bundle ZIP.
• FFmpeg source and build-information archives are included for the clean LGPL audio runtime.

See the full changelog for details.