Changelog
v0.0.107
Improvements
- Faster performance on Windows: Windows x64 builds now enable AVX, AVX2, FMA, and F16C CPU optimizations, delivering noticeably better speed on compatible hardware.
v0.0.106
No notable changes in this release.
v0.0.105
Improvements
-
Sound effects settings are now more reliable. If you had never changed a sound effect option before, it will still behave correctly with sensible defaults — no more unexpected silence or missing error sounds.
-
Whisper backend selector simplified. The dropdown for choosing your speech‑to‑text engine (Auto, CPU, GPU) no longer shows hint labels like “Recommended” or “Accelerated”, making the list cleaner and easier to scan at a glance. The icons have also been updated for a more consistent look.
-
Settings dropdowns are more compact. The width of dropdown menus in the General and Whisper Backend sections has been reduced, giving the settings panel a tighter, more balanced layout.
v0.0.104
New Features
- GPU-accelerated transcription — Speech recognition can now use your graphics card (GPU) for faster processing. The app supports Vulkan on Windows and Linux, and Metal on macOS, giving you accelerated performance when a compatible GPU is available.
- Simplified backend selection — The speech recognition backend setting has been streamlined. Instead of separate options for CUDA, Vulkan, Metal, and CPU, you now simply choose Auto, CPU, or GPU. The app automatically detects what acceleration your system supports, so you don't need to worry about the technical details.
Bug Fixes
- Fixed an issue where the recording start time could be reported incorrectly in certain scenarios.
v0.0.103
Improvements
- Recording status overlay now stays visible more reliably on macOS. The overlay window's display level has been adjusted so it appears above most windows, including full-screen apps, making it easier to see when a recording is active. The overlay still won't interfere with your work — it ignores mouse clicks and doesn't steal focus.
v0.0.102
Improvements
- macOS recording overlay overhaul: The recording status indicator now uses a proper system panel instead of custom window manipulation. This means the overlay reliably stays on top of other windows, works correctly in full-screen spaces, and no longer interferes with your keyboard focus or mouse clicks.
- Cleaner overlay appearance: The overlay window now has proper transparency and no drop shadow, for a more polished look during recording.
New Features
- Custom build identifier: If your app was built with a custom build environment variable (e.g., "staging"), a badge now appears next to the version number in Settings so you can easily tell which build you're running.
Internal Changes
- Removed several obsolete configuration files (
.npmrc,.gitattributes,AGENTS.md) that were no longer needed.
v0.0.101
Improvements
- Settings page redesigned – The Settings page has been reorganized into clear sections (General, Input Device, Advanced, About) making it much easier to find and adjust the options you need.
- Audio input device handling – You can now choose whether to always use your system's default microphone or manually select and prioritize specific devices. A new "Use system default" toggle makes switching between the two modes quick.
- Individual sound effect controls – Instead of a single "disable sound effects" switch, you can now turn each sound on or off independently: start recording sound, stop recording sound, error sound, and no-speech (silent audio) sound. Find them under the new "Sound effects" section in Advanced settings.
- New start/stop recording sounds – You can optionally enable a sound that plays when recording begins and another when recording ends. Both are disabled by default.
- Sound feedback timing improved – The start recording sound now plays after recording actually starts, and the stop recording sound plays after audio processing is complete, making the feedback more accurate.
- Logo shown in About section – The app logo now appears in the Settings page's About section alongside the version number.
- App name and version shown dynamically – The Settings page now displays the actual application name and version, correctly reflecting the current build instead of using fixed text.
- "Whisper Hardware" label – The speech-recognition backend setting has been renamed from "Whisper backend" to "Whisper Hardware" for clarity.
- New profiles default to English – When adding a new profile, the language is automatically set to English, saving you an extra step.
- Profile deletion navigates to overview – Deleting a profile now takes you back to the overview page, providing a clearer flow.
Bug Fixes
- Fixed sound playing before recording starts – The start-recording sound no longer plays before the recording has actually begun, ensuring accurate audio feedback.
- Simplified config directory setup – Removed the old migration logic that moved configuration from a legacy folder. The app now creates its configuration directory directly without relying on deprecated paths.
v0.0.100
Improvements
- Updated application icons and branding. Icons are now automatically generated during each build, ensuring crisp, consistent logos across all platforms. The development build's logo color has been updated from red to amber to better match the brand identity.
- Config directory has moved to a standard location. Your settings are now stored in the OS‑standard app config directory. Old settings from
~/.oiperor~/.oiper.devare automatically migrated on first launch — no action needed. - Downloads now show progress and can be cancelled. When the app downloads files (such as whisper models), you'll see real‑time progress updates, and you can cancel an in‑progress download if needed.
- Refined recording pipeline. The app now uses a single, more reliable vocal‑detection check to decide whether a recording should be processed, making the flow faster and more consistent.
Bug Fixes
- A "stop" sound now plays when a recording is skipped. If the app detects no voice in a recording and discards it, you'll hear a subtle sound so you know what happened.
- Fixed sound file paths on macOS. Sound effects now resolve correctly on macOS, preventing missing‑sound errors.
- Release workflows now generate logo assets correctly. Logo generation is now a proper step in the release pipeline, ensuring release builds include the correct icons.
Breaking Changes
- The configuration directory has changed. Your settings are now stored in the platform‑standard app config directory (e.g.,
~/Library/Application Support/com.oiper.desktop/on macOS,~/.config/oiper-desktop/on Linux). If you were relying on the old~/.oiperor~/.oiper.devpaths, the app will automatically migrate your data on first launch. No user action is required.
v0.0.99
New Features
- Choose and prioritize your audio input devices. You can now see all your connected microphones and audio input devices directly in Settings under a new "Audio Input" section. Drag to reorder them — the first working device in your list will be used for recording. Devices that are currently disconnected are shown with a "Not connected" badge but stay in your list so they're ready when you plug them back in. A "System Default" option is always available if you'd rather let your operating system decide.
v0.0.98
Bug Fixes
- Fixed audio playback in history details – Audio recordings can now be played back and revealed in your file explorer reliably, as the app now looks for them in the correct location.
v0.0.97
No notable changes in this release.
v0.0.96
Bug Fixes
- Fixed an issue where blank or silent audio segments could sometimes produce unexpected results. The app now properly handles these cases and skips them cleanly, preventing garbled or empty text from appearing in your transcriptions.
v0.0.95
Improvements
- Refreshed application icons – The app icons have been updated with new colors and improved visuals across all platforms (Windows, macOS, Linux, iOS, and Android). The system tray icon and development-mode overlay color have been updated too for a more polished look.
Bug Fixes
- Fixed macOS hotkey reliability – Resolved an issue where keyboard shortcuts could behave unreliably on macOS, particularly during recordings. Hotkey detection now works correctly on Mac, ensuring your shortcuts are always recognized when they should be.
v0.0.94
Breaking Changes
- The experimental
experimental_use_grabbersetting has been replaced by a newbypass_hotkeyoption. If you previously setexperimental_use_grabbertotrue, you now need to setbypass_hotkeytofalse(the default) to keep the same behavior. The new option's purpose is inverted: enablingbypass_hotkeytells the app to stop suppressing system hotkeys, instead of the old behavior where enabling the grabber would actively suppress them. The setting label in the user interface has been updated to "Bypass hotkey (stops suppressing hotkeys, restarts app)" to reflect this. Any existingexperimental_use_grabberconfiguration is no longer recognized and will be ignored.
v0.0.93
New Features
- Experimental key grabber for better hotkey handling: You can now enable an experimental "key grabber" mode in Settings that prevents registered hotkeys from reaching other applications. This means when you use a keyboard shortcut for OiPer, it won't also trigger actions in whatever other app is focused. To try it, go to Settings → Experimental and toggle "Use key grabber". The app will restart to apply the change. (Note: this is an experimental feature — please let us know if you run into any issues.)
Improvements
- Restart capability added: The app can now restart itself automatically when needed, for example after changing the experimental key grabber setting.
v0.0.92
Bug Fixes
- Fixed tray icon not showing correctly on Windows in development mode. The icon now uses the proper image format, ensuring it displays as expected when running a development build on Windows.
- Fixed tray icon appearance on macOS. The icon is no longer treated as a system template, so it keeps its intended look regardless of whether you’re using light or dark mode.
v0.0.91
New Features
- Dedicated system tray icon – The app now shows its own icon in the system tray (notification area). Clicking the tray icon opens or closes the main window, making it easier to quickly access the app without searching for it in your taskbar or dock.
Improvements
- Refreshed application icon – The app icon has been redesigned: it now uses a square shape with rounded corners and a darker background, giving it a more modern and polished look. You'll see the new icon on your desktop, taskbar, dock, and in your app launcher.
- Cleaner tray icon – The icon that appears in the system tray has been updated to match the new design, ensuring a consistent visual experience wherever you interact with the app.
- Smoother tray interactions – The way the app behaves when you click the tray icon has been improved, making show/hide actions more reliable.
v0.0.90
Improvements
- Updated app icon and logo. The app has a fresh new look with redesigned icons across all platforms — including the taskbar, dock, mobile, and browser tab — giving Oiper a more polished and modern appearance.
v0.0.89
Improvements
- Updated app icons and logo. The application icons across all platforms (Windows, macOS, Android, iOS) and the web logo have been refreshed with a new look — the background color has changed from green to black, giving the app a sleeker appearance.
v0.0.88
Improvements
- Toast notifications now appear with richer colors, making them easier to notice and distinguish at a glance.
New Features
- On macOS and Linux, checking for updates now clearly informs you that automatic update checking is currently supported on Windows only.
Bug Fixes
- Updated the application icon and logo throughout the app — the icon now features a fresh, simplified design with new colors.