JunJulJul 5, 20260.2.5Global dictation hotkeys now work right after installing the deb/rpm, with no input-group setup or re-login neededFixes deb install conflict with other apps shipping ONNX Runtime0.2.4History entries now show the real audio duration. deb and rpm installs auto-update from this version on.0.2.3First Linux release: real-time dictation with global hotkeys, file transcription, AI modes, local Whisper/Parakeet models, and cloud transcription.Jul 16, 20260.3.3Fixes missing start and stop audio cues in Linux packages.0.3.2Adds Pro settings sync across devices, improves sync reliability, and fixes background memory use from the hidden overlay.Jul 17, 20260.3.5Fixes text insertion typing "2442" instead of the transcript on systems with ydotool 0.1.x (Ubuntu's packaged version).Jul 20, 20260.3.8Mouse side buttons (X1/X2) can now be used as dictation shortcutsF13-F24 keys now work as shortcutsFixed hotkeys stopping to work after plugging in input devices or recording a shortcutElevenLabs: real-time streaming transcription with your own API keySoniox: context (dictionary) settings to boost recognition of your termsGPT-4o transcription prompt settingsPrompt settings sync across your devicesAlerts you when the microphone stays silent while recordingGeneral bug fixes and stability improvementsAugJul 27, 20260.3.11See live transcription as you speak with ElevenLabs ScribeDictate with AssemblyAI Universal 3.5 Pro, built for noisy audioFile transcription now produces accurate subtitle timings with local models and your own API keysDictated text stays out of clipboard history and cloud clipboard syncFixes: upgrade prompt opens in front of other windows, no stray Enter after a failed dictation, multi-track files with a silent audio track transcribe correctlyAug 3, 20260.3.13Fixed: the app opened to a blank white window on Arch, Fedora and other rolling-release distros (AppImage)Rendering now recovers automatically if the system graphics stack fails to startRedesigned grouped settings sheets; language and key-term options for GPT-4o and SonioxNew installs start with ElevenLabs Scribe by defaultUpdate checks now appear quietly in the sidebar and trayCrash reporting on Linux and Windows for faster fixesAug 7, 20260.3.15Bug fixes and stability improvementsAug 11, 20260.3.17Bug fixes and stability improvementsAug 27, 20260.3.19New Gemini 3.5 Transcribe dictation model, auto-detects 85+ languagesAI providers: OpenAI Responses API support (base URL ending in /responses)Fixed a false "File too large" warning for video files and local modelsBug fixes and stability improvementsAug 29, 20260.3.20Option to copy dictated text to the clipboardBug fixes and stability improvementsSepSep 5, 20260.3.21Fixed Quick Send when ending dictation with Enter on Wayland.Added setup guidance when Linux keyboard permissions are missing.Added MAI-Transcribe 2 with language, keyterm, and transcript cleanup settings.Added live English dictation with Cartesia ink-2 using your own API key.Sep 6, 20260.3.23Added Ubuntu 20.04 support to the Linux AppImage.0.3.22Fixed the recording shortcut not responding after waking from sleepSep 7, 20260.3.25Fixed Wi-Fi turning off at startup due to incorrect virtual radio switch initialization.0.3.24Fixed a blank window when launching the AppImage on Ubuntu 20.04.Sep 15, 20260.3.26Fixed dictated text not pasting on Wayland without a paste helper installedAdded an optional language field to OpenAI-compatible providersVarious bug fixes and stability improvementsSep 21, 20260.3.30Fixed Ctrl+Z in a terminal freezing keyboard and mouse inputFixed Ctrl staying pressed after resuming or unplugging a keyboardVarious bug fixes and stability improvementsSep 24, 20260.4.0New Home screen with your dictation streak, pace and statsShare cards that show off your dictation progressFixed downloading local models in local-only modeFixed local Whisper models crashing on older processorsFixed the AppImage not starting on Fedora and other distrosVarious bug fixes and stability improvementsOctSep 30, 20261.0.1Fixed the shortcut sometimes not working on WaylandBrowse provider models when picking an AI modelEach mode can now use its own dictation modelVarious bug fixes and stability improvementsOct 1, 20261.0.3Italian and Brazilian Portuguese interface languagesFixed getting stuck in the onboarding flowVarious bug fixes and stability improvementsOct 5, 20261.0.5New Capsule recording display style, available with ProSmart Paste now pastes into the Limux terminal on X111.0.4Smart Paste now pastes into GNOME Terminal and several other terminalsOct 9, 20261.0.7Fixed virtual input devices staying after a keyboard is unpluggedFixed a resource leak each time an input device is reconnected1.0.6Fixed custom shortcut recording in Wayland sessionsFixed a startup crash on computers with several usersShortcut labels now show Super instead of WinOct 10, 20261.0.8Fixed the screen locking during a long dictation on GNOMEFixed the recording overlay and text insertion under swayFixed a crash at startup when no display can be openedOct 11, 20261.0.10Fixed the app not starting after installation from the rpm package1.0.9New Local Model Cleanup removes filler words and fixes punctuationTurn it on in Settings > Behavior > Local Model CleanupOct 12, 20261.0.11Various bug fixes and stability improvementsOct 22, 20261.2.0New AI Models screen keeps dictation and LLM keys in one placeUpdated the suggested model lists for every providerLocal models now open a quick test right after downloadVarious bug fixes and stability improvementsNovNov 3, 20261.3.0Cloud dictation adds Fun-ASR Flash, Whisper Large v3 and Gemini LiveNew shortcut pastes your last transcription into any appDeepSeek and OpenCode Go added as AI processing providersReal-time Nova-3 and Cartesia Ink now show text as you speakMute while recording now silences system audio on LinuxVarious bug fixes and stability improvementsNov 5, 20261.4.0New Mobile Dictation lets you dictate from your phoneSystem audio can now be lowered instead of muted while you dictateVarious bug fixes and stability improvementsNov 8, 20261.4.1Local models stay loaded for 2 hours, so dictation starts fasterVoxtral Small Transcribe 3 is available with your own Mistral keyVarious bug fixes and stability improvementsNov 13, 20261.4.2Scribe v2 Medical and new Scribe v3 options in Spokenly CloudMicrosoft Foundry dictation is available with your own keyAPI keys can now be named and renamed in AI ModelsVarious bug fixes and stability improvementsNov 16, 20261.5.2A cleaner, more consistent interface across the app1.5.0Local models now run on your GPU through VulkanLocal models fall back to the CPU if the GPU driver failsParakeet Unified is now available with live text previewRe-download your Parakeet model to get the GPU versionAI processing with Claude now caches prompts to save costsVarious bug fixes and stability improvementsNov 17, 20261.5.3Fixed the microphone in the AppImage on Arch, Fedora and other distributionsNov 18, 20261.5.4Various bug fixes and stability improvements
1.5.4
Latest
Various bug fixes and stability improvements
1.5.3
Fixed the microphone in the AppImage on Arch, Fedora and other distributions
1.5.2
A cleaner, more consistent interface across the app
1.5.0
Local models now run on your GPU through Vulkan
Local models fall back to the CPU if the GPU driver fails
Parakeet Unified is now available with live text preview
Re-download your Parakeet model to get the GPU version
AI processing with Claude now caches prompts to save costs
Various bug fixes and stability improvements
1.4.2
Scribe v2 Medical and new Scribe v3 options in Spokenly Cloud
Microsoft Foundry dictation is available with your own key
API keys can now be named and renamed in AI Models
Various bug fixes and stability improvements
1.4.1
Local models stay loaded for 2 hours, so dictation starts faster
Voxtral Small Transcribe 3 is available with your own Mistral key
Various bug fixes and stability improvements
1.4.0
New Mobile Dictation lets you dictate from your phone
System audio can now be lowered instead of muted while you dictate
Various bug fixes and stability improvements
1.3.0
Cloud dictation adds Fun-ASR Flash, Whisper Large v3 and Gemini Live
New shortcut pastes your last transcription into any app
DeepSeek and OpenCode Go added as AI processing providers
Real-time Nova-3 and Cartesia Ink now show text as you speak
Mute while recording now silences system audio on Linux
Various bug fixes and stability improvements
1.2.0
New AI Models screen keeps dictation and LLM keys in one place
Updated the suggested model lists for every provider
Local models now open a quick test right after download
Various bug fixes and stability improvements
1.0.11
Various bug fixes and stability improvements
1.0.10
Fixed the app not starting after installation from the rpm package
1.0.9
New Local Model Cleanup removes filler words and fixes punctuation
Turn it on in Settings > Behavior > Local Model Cleanup
1.0.8
Fixed the screen locking during a long dictation on GNOME
Fixed the recording overlay and text insertion under sway
Fixed a crash at startup when no display can be opened
1.0.7
Fixed virtual input devices staying after a keyboard is unplugged
Fixed a resource leak each time an input device is reconnected
1.0.6
Fixed custom shortcut recording in Wayland sessions
Fixed a startup crash on computers with several users
Shortcut labels now show Super instead of Win
1.0.5
New Capsule recording display style, available with Pro
Smart Paste now pastes into the Limux terminal on X11
1.0.4
Smart Paste now pastes into GNOME Terminal and several other terminals
1.0.3
Italian and Brazilian Portuguese interface languages
Fixed getting stuck in the onboarding flow
Various bug fixes and stability improvements
1.0.1
Fixed the shortcut sometimes not working on Wayland
Browse provider models when picking an AI model
Each mode can now use its own dictation model
Various bug fixes and stability improvements
0.4.0
New Home screen with your dictation streak, pace and stats
Share cards that show off your dictation progress
Fixed downloading local models in local-only mode
Fixed local Whisper models crashing on older processors
Fixed the AppImage not starting on Fedora and other distros
Various bug fixes and stability improvements
0.3.30
Fixed Ctrl+Z in a terminal freezing keyboard and mouse input
Fixed Ctrl staying pressed after resuming or unplugging a keyboard
Various bug fixes and stability improvements
0.3.26
Fixed dictated text not pasting on Wayland without a paste helper installed
Added an optional language field to OpenAI-compatible providers
Various bug fixes and stability improvements
0.3.25
Fixed Wi-Fi turning off at startup due to incorrect virtual radio switch initialization.
0.3.24
Fixed a blank window when launching the AppImage on Ubuntu 20.04.
0.3.23
Added Ubuntu 20.04 support to the Linux AppImage.
0.3.22
Fixed the recording shortcut not responding after waking from sleep
0.3.21
Fixed Quick Send when ending dictation with Enter on Wayland.
Added setup guidance when Linux keyboard permissions are missing.
Added MAI-Transcribe 2 with language, keyterm, and transcript cleanup settings.
Added live English dictation with Cartesia ink-2 using your own API key.
0.3.20
Option to copy dictated text to the clipboard
Bug fixes and stability improvements
0.3.19
New Gemini 3.5 Transcribe dictation model, auto-detects 85+ languages
AI providers: OpenAI Responses API support (base URL ending in /responses)
Fixed a false "File too large" warning for video files and local models
Bug fixes and stability improvements
0.3.17
Bug fixes and stability improvements
0.3.15
Bug fixes and stability improvements
0.3.13
Fixed: the app opened to a blank white window on Arch, Fedora and other rolling-release distros (AppImage)
Rendering now recovers automatically if the system graphics stack fails to start
Redesigned grouped settings sheets; language and key-term options for GPT-4o and Soniox
New installs start with ElevenLabs Scribe by default
Update checks now appear quietly in the sidebar and tray
Crash reporting on Linux and Windows for faster fixes
0.3.11
See live transcription as you speak with ElevenLabs Scribe
Dictate with AssemblyAI Universal 3.5 Pro, built for noisy audio
File transcription now produces accurate subtitle timings with local models and your own API keys
Dictated text stays out of clipboard history and cloud clipboard sync
Fixes: upgrade prompt opens in front of other windows, no stray Enter after a failed dictation, multi-track files with a silent audio track transcribe correctly
0.3.8
Mouse side buttons (X1/X2) can now be used as dictation shortcuts
F13-F24 keys now work as shortcuts
Fixed hotkeys stopping to work after plugging in input devices or recording a shortcut
ElevenLabs: real-time streaming transcription with your own API key
Soniox: context (dictionary) settings to boost recognition of your terms
GPT-4o transcription prompt settings
Prompt settings sync across your devices
Alerts you when the microphone stays silent while recording
General bug fixes and stability improvements
0.3.5
Fixes text insertion typing "2442" instead of the transcript on systems with ydotool 0.1.x (Ubuntu's packaged version).
0.3.3
Fixes missing start and stop audio cues in Linux packages.
0.3.2
Adds Pro settings sync across devices, improves sync reliability, and fixes background memory use from the hidden overlay.
0.2.5
Global dictation hotkeys now work right after installing the deb/rpm, with no input-group setup or re-login needed
Fixes deb install conflict with other apps shipping ONNX Runtime
0.2.4
History entries now show the real audio duration. deb and rpm installs auto-update from this version on.
0.2.3
First Linux release: real-time dictation with global hotkeys, file transcription, AI modes, local Whisper/Parakeet models, and cloud transcription.