macOS Releases

July 29, 2026

• -New GPT Transcribe online model for Pro users
• -Scripts can now switch modes and control output with new directives
• -Fixed Smart Paste inserting stale text in iPhone Mirroring
• -More reliable text insertion in Chrome-based apps

July 28, 2026

• -Transcribe channels separately from the CLI with --channels and --names
• -Improved file transcription through the ElevenLabs API
• -Fixed silence trimming corrupting stereo and high sample rate audio

July 27, 2026

• -Keep dictating when a transcription provider goes down
• -Identify speakers in File Transcription with ElevenLabs Scribe
• -Improved AssemblyAI support with realtime dictation and keyterms

July 24, 2026

• -Insert dictated text from a floating panel when no field is focused
• -See live transcription as you speak with ElevenLabs Scribe
• -Dictate with AssemblyAI Universal 3.5 Pro, built for noisy audio
• -Improved Getting Started flow and general reliability

July 20, 2026

• -Transcribe stereo channels separately for clearer speaker separation
• -Sync GPT-4o prompts and Soniox context across your devices
• -Get a helpful alert when no microphone audio is detected
• -Improved AI processing feedback and transcription reliability

July 17, 2026

• -Fixed the Notch recording indicator not appearing during dictation

July 16, 2026

• -Switch modes by voice with trigger words
• -Improved local model download reliability

July 15, 2026

• -Sync selected settings across devices with Spokenly Pro
• -Access recent transcriptions directly from the menu bar
• -View paste destinations in History and search by app
• -Customize Parakeet Unified latency for speed and energy use
• -Improved overlay, trackpad, and subscription reliability

July 5, 2026

• -Fixed a crash triggered by trackpad gestures
• -Minor bug fixes and improvements

July 3, 2026

• -Improved trackpad send swipe reliability

July 3, 2026

• -New trackpad and mouse triggers for dictation shortcuts
• -Trackpad gestures now support swipe-to-send
• -Per-mode Smart Capitalization controls for custom Modes
• -Reliability fixes for triggers and local transcription

July 1, 2026

• -Smart Capitalization now respects your Word Replacements
• -Improved recovery for interrupted File Transcriptions
• -Added a separate Default Mode shortcut setting for simpler setups

June 29, 2026

• -Improved punctuation handling for WebView

June 28, 2026

• -Added Amazon Bedrock as a new AI provider
• -Added Inception as a new AI provider
• -Improved punctuation handling during text insertion
• -Added JSON export and import for Word Replacements

June 19, 2026

• -Re-transcription now respects your Word Replacements
• -More reliable handling of trailing punctuation
• -Word Replacements shortcut no longer opens its window repeatedly
• -Sarvam API model now offers a configurable output mode

June 18, 2026

• -Modes naming is now used across the entire app
• -History now details every stage of the processing pipeline, with diffs
• -Modes can now be activated by multiple apps, not just a single one
• -Modes can now activate automatically on specific websites
• -Smart Capitalization and Smart Spacing are now separate toggles in Settings
• -Added NVIDIA Parakeet Unified, a new fully on-device transcription model

June 7, 2026

• -Added Microsoft MAI-Transcribe 1.5 model for dictation
• -Fixed text insertion not working in some cases
• -Fixed duplicate microphones in the input device list
• -Added OpenRouter speech-to-text support with your own API keys
• -Fixed Cohere model returning only part of the transcript

June 4, 2026

• -New on-device Nemotron 3.5 model for multilingual dictation
• -New on-device Cohere model for dictation
• -Improved overall app stability and reliability

June 3, 2026

• -Invite friends to earn free Pro
• -Added support for the Corti real-time transcription API
• -Stability and reliability improvements

May 28, 2026

• -Interrupted streaming dictations now keep their audio in History for retranscription

May 21, 2026

• -Send with Return key now works in any app, configurable in Settings
• -Local models now produce more reliable punctuation at the end of dictation
• -Fixed occasional freezes in terminals with very large text buffers

May 13, 2026

• -Improved accessibility for VoiceOver
• -More reliable cursor context for web fields
• -Experimental BYOK support for gpt-realtime-whisper

May 12, 2026

• -Richer transcription metadata in History details
• -Improved History performance and stability
• -Deepgram now supports language hints
• -Various bug fixes and stability improvements

May 10, 2026

• -Smart Spacing now handles capitalization and punctuation more reliably
• -More reliable cursor context capture in Electron-based apps

May 9, 2026

• -New: Smart Paragraphs splits long dictation into readable paragraphs
• -New: Smart Capitalization for natural mid-sentence dictation
• -New: Deepgram Flux model for low-latency streaming transcription
• -Modes can now use the active browser URL as context
• -Various stability improvements and bug fixes

April 25, 2026

• -Spokenly CLI now supports JSON output
• -Fewer cases of Apple Music opening during dictation
• -Parakeet model updated. Please redownload it to keep using on-device transcription
• -Various stability improvements

April 20, 2026

• -Added xAI Grok transcription model option
• -Added AppleScript paste mode compatible with any keyboard layout
• -Bash scripts now receive context via env vars
• -Fixed Apple Speech Analyzer truncating the last words of realtime transcripts
• -MCP integration now supports AI Instructions for transcribed text
• -Various stability improvements and bug fixes

April 6, 2026

• -Spokenly CLI is now available
• -MCP now includes a tool for transcribing files
• -Local speaker identification powered by Pyannote
• -Fixed shortcuts firing incorrectly after system wake

April 2, 2026

• -Soniox API key support removed to comply with Soniox Terms of Service
• -Fixed smart paste resetting input source for CJK languages
• -General bug fixes and improvements

March 31, 2026

• -Real-time transcription with Apple Speech Analyzer
• -Added support for Opus and OGG audio formats
• -Fixed a crash on Intel Macs

March 30, 2026

• -Improved voice activity detection for quieter speech
• -VAD trimming can now be toggled off in settings
• -Improved AssemblyAI transcription support

March 30, 2026

• -Audio is now trimmed before sending to cloud APIs for faster results
• -Fixed an issue with invalid text ranges in accessibility context
• -Improved diagnostic logging for easier troubleshooting

March 24, 2026

• -Smart spacing automatically adds spaces around dictated text
• -Cursor context limited to 2000 characters to reduce token usage
• -Duplicate modes from the right-click menu
• -Various bug fixes and stability improvements

March 18, 2026

• -Faster recording start and finish with improved animations
• -LM Studio AI provider for local models
• -Automatic media pausing during recording
• -Word Replacements editing fix and stability improvements

March 16, 2026

• -MCP server status page viewable in browser
• -Voice now plays for every Agent Mode question
• -Added "Use system default" microphone setting
• -Automatic recovery of interrupted recordings

March 12, 2026

• -Voice for AI Agents: use Spokenly as MCP server in Claude, Codex, and Cursor
• -Better microphone handling with external displays
• -Recording popup stays above all windows on every space
• -Cmd+Q keeps the app running in the menu bar

March 8, 2026

• -Escape key now properly dismisses the recording overlay
• -Fixed freezing issues with real-time streaming models
• -Resolved crashes when using local transcription models
• -Soniox API now supports Japan and Europe region selection

March 6, 2026

• -Fixed crashes when getting cursor context for LLM
• -Scroll indicators now display correctly in ScrollViews
• -New Accessibility API typing mode for custom keyboard layouts

March 5, 2026

• -Improved Agent mode stability and reliability
• -Enhanced Volcengine speech recognition support

March 4, 2026

• -Bug fixes and stability improvements