Changelog
Dicta updates
Latest release: v1.5.11
1.5.11 — February 21, 2026
Added
- QTA file support in import/transcription flow.
Fixed
- Fixed onboarding microphone monitoring race condition that could leave microphone capture active after leaving setup.
- Added defensive recorder cleanup to ensure CoreAudio engine is stopped during deallocation.
1.5.10 — February 20, 2026
Fixed
- Removed misleading "Already subscribed? Sign in" link on the subscription screen when you're already signed in.
1.5.9 — February 19, 2026
Fixed
- Fixed transcription slowdown from 1.5.8 — short recordings (2–12 s) were taking 20–48 s instead of 2–3 s.
1.5.8 — February 19, 2026
Improved
- 3–5× faster transcription for short recordings — eliminated audio repetition that previously caused ~15 s latency for typical 5 s dictations.
Fixed
- App no longer steals focus from other windows when launched at login.
- Opening Dicta via Dock or Cmd+Tab while onboarding is in progress now correctly shows the setup wizard.
- Clicking the menu bar icon during onboarding now correctly shows the setup wizard.
1.5.6 — February 18, 2026
Changed
- Unified Account section in Settings — subscription plan, billing actions, and device management are now combined in one place.
- Subscription status text improved: color-coded status with inline date display (e.g., "Trial — ends Feb 21").
- Trial expiry and payment issue banners can now be dismissed with an × button.
- Permission banners are consistently dismissable.
Fixed
- Model first-time setup now shows elapsed time so you know compilation is still in progress.
1.5.5 — February 18, 2026
Changed
- "See plans" now opens your account page already signed in, so you can choose a plan without re-entering your email.
- All warning banners now show their action button inline for better visibility and a cleaner look.
Fixed
- Trial expiry CTA was easy to miss. Now it's an inline action on the banner title row.
1.5.4 — February 17, 2026
Changed
- Microphone permission banner now stays visible whenever access is missing.
- Settings now prioritizes Permissions at the top when attention is required.
- Permission action buttons are now more prominent in Settings.
Fixed
- Replaced confusing "banner above" guidance with direct path: Settings > Permissions.
1.5.3 — February 17, 2026
Changed
- Onboarding no longer blocks on model download. Setup continues while the model prepares in the background.
- Try Recording flow is simpler with clear model status and continue options.
- Code verification is smoother with clearer inline flow and a shorter resend cooldown.
- Permissions copy now better explains Accessibility as recommended for full experience.
Fixed
- Reopening setup wizard no longer forces account creation for users already signed in.
- Removed fragile "Open Mail" deep-link behavior in onboarding.
1.5.2 — February 17, 2026
New: Meeting Mode
- Record meetings and long sessions up to 8 hours.
- Audio saved to disk — transcribed after you stop.
- Switch between Dictation and Meeting modes with one click.
- Crash recovery: interrupted recordings can be recovered on next launch.
Account Management
- New Account section in Settings: see your email, manage devices, sign out.
- View and disconnect devices from your account.
Improved
- Streamlined onboarding: fewer steps, permissions and mic test merged.
- Better error messages for corrupted model files.
- Settings sections reordered: Input and Behavior first.
1.5.1 — February 15, 2026
Added
- Subscription status updates faster after payment — no more waiting.
- Sign-in code auto-submits when you type the 6th digit.
- New "More than dictation" feature discovery during onboarding.
Improved
- Clearer, jargon-free recording mode setup in onboarding.
- Settings shows specific messages based on your account status.
- Onboarding no longer gets stuck after account transfer.
- Better window sizing for onboarding wizard.
Fixed
- Empty transcription in Try It Now demo now resets properly.
1.5.0 — February 14, 2026
Added
- Email-first trial onboarding: new users sign up with email during onboarding for a smoother 14-day trial experience.
- Payment issue banner: proactive in-app warning when a payment problem needs attention.
- Migration support: existing users get a guided transition to the new sign-in flow.
Changed
- Architecture overhaul: recording state machine, dedicated permission manager, API client extraction, and view decomposition for a cleaner codebase.
- Shared checkout logic across activation, trial-expired, and settings views.
1.4.8 — February 13, 2026
Added
- Update notification in main window: a blue "Update" pill now appears when a new version is available. Click it to download and install.
- History multi-select: use Cmd+Click, Shift+Click, and Cmd+A to select multiple transcriptions. Batch delete with confirmation and undo support.
Improved
- Sparkle update diagnostics for troubleshooting auto-update issues.
1.4.7 — February 13, 2026
Fixed
- Intel crash on stop recording: fixed crash on Intel Macs when stopping recording.
- Hotkey reliability after stop: added event-tap health checks with automatic recovery.
- Silent hotkey no-op during transcribing: ⌥Space now shows immediate busy feedback instead of silently doing nothing.
- Recording start hang mitigation: microphone startup with timeout, automatic fallback to system default.
Improved
- Insertion fallback: unified insertion path falls back to clipboard when automatic paste fails.
- Insertion diagnostics: text inserter surfaces technical paste simulation errors for better recovery.
1.4.6 — February 12, 2026
Added
- Universal binary: runs natively on both Apple Silicon (M1/M2/M3/M4) and Intel Macs.
- Drag & drop from History: drag transcriptions from History browser or the latest preview card into any app.
- File access recovery: clear error messages and a "Choose file…" button when macOS blocks file access.
Improved
- History browser: unified glass background, refined column divider.
- File picker: unified selection behavior across DMG and App Store builds.
1.4.5 — February 12, 2026
Changed
- Import field: contextual action button — upload icon when empty, send when ready, disabled state for invalid input.
- Import field icons: leading icon adapts to input type (link, document, or hidden).
Added
- CAF format support: Core Audio Format (.caf) files can now be imported and transcribed.
Improved
- File picker consistency: unified content type list across menu import and main window picker.
1.4.4 — February 11, 2026
Improved
- History browser: per-column backgrounds with glass sidebar and subtle gradient detail panel.
- Activation code field: cleaner digit-only input handling.
- Onboarding: App Store permissions copy now focused on text insertion.
- Settings: improved sign-in flow opens main window instead of clearing session.
1.4.3 — February 11, 2026
Changed
- Standalone main window replaces popover for better Dock and Cmd+Tab integration.
- Glass visual effects across main window, settings, and history.
- Settings now opens via native SwiftUI Settings scene.
- Centralized activation form state for faster keystroke rendering.
Fixed
- Swift 6 actor-isolation warnings for recording limit constants.
1.4.2 — February 10, 2026
Improved
- Postponed updates now auto-install when Dicta becomes idle instead of waiting indefinitely.
- Background update check runs shortly after launch for faster update delivery.
1.4.1 — February 10, 2026
Fixed
- Audio device buffer allocation that could misdetect input channels.
- Stale device IDs after Continuity microphone reconnects.
Improved
- Weak signal guidance now suggests switching to iPhone mic when detected.
- Automatic fallback to system default if preferred microphone fails.
- Input format validation before starting audio engine.
Changed
- Import field now shows a privacy note about local transcription.
1.4.0 — February 10, 2026
Added
- Unified design system with consistent visual tokens across all views.
- Redesigned onboarding wizard with improved layout and flow.
- History sidebar now shows entry count and total transcription duration.
- Continuity guidance banner for multi-device setup help.
Changed
- Refreshed interface across main view, activation, settings, and trial screens.
- Permission banners now use a consistent callout component.
- History respects Reduce Motion accessibility preference.
Fixed
- Prevented duplicate sign-in requests during activation.
- Improved Sparkle update observer reliability.
1.3.3 — February 9, 2026
Changed
- Whisper long-form defaults — Background streaming now disables cross-chunk prompt conditioning by default to reduce premature endings on long dictations.
Fixed
- Transcription tail clipping — Chunk merge now falls back to text merge when word timestamps are incomplete, preventing dropped endings.
- Low-volume final tail loss — Final chunk is now processed even when RMS is below silence threshold; silence skipping remains for non-final chunks.
1.3.2 — February 9, 2026
Fixed
- Settings not opening from gear — Settings now opens as a popover with apple glass effect instead of a separate window.
- Settings from context menu and app menu — Right-click and menu bar "Settings" now work correctly.
- Icon spacing — Tightened icon button spacing to match visual density.
- Layout symmetry — Equalized gaps above and below the record button.
1.3.1 — February 9, 2026
Changed
- Esc closes popover — Press Escape to dismiss the main popover, matching standard macOS behavior.
- Default action buttons — Primary buttons now respond to Enter/Return key across all dialogs.
- Semantic colors — UI adapts correctly to Light and Dark mode using system color tokens.
- Context menu polish — Improved right-click menus with consistent styling and keyboard shortcuts.
1.3.0 — February 9, 2026
Added
- Free 14-day trial — Download and use Dicta free for 14 days, no email or credit card required.
- Post-trial paywall — Dedicated pricing view with plan picker and Subscribe CTA when trial expires.
- Trial banner visible from 5 days remaining with actionable Subscribe button.
- Settings subscription destination — Deep-link to subscription section from trial banner and other views.
Changed
- HIG audit — Drag-and-drop file import, fullscreen History, Cmd+W close, context menu, improved hover states.
- Unified Settings window — Gear opens native macOS Settings window with reliable focus handling.
Fixed
- Settings not opening when clicking gear icon in popover.
- Help menu build error due to CommandsBuilder item limit.
1.2.8 — February 8, 2026
Changed
- Whisper long-form parity (mixed) — Streaming transcription now carries prompt context between windows, locks auto-detected language after the first successful chunk, and advances seek using word timestamps to match upstream Whisper behavior more closely.
- Recording limits unified — Recording timer and audio buffer limits now use shared constants with a 60-minute cap for long dictations.
Fixed
- Long dictation cut-offs — Final chunk clipping now avoids trimming short tails, and chunk merge preserves text when word timestamps are missing in any chunk.
- Stop reason diagnostics — Recording stops now log explicit reason and duration (`manual`, `escape`, `max_duration`, `error`) for easier debugging.
1.2.7 — February 7, 2026
Changed
- Settings open reliability — Opening Settings now consistently activates Dicta and dispatches window presentation from menu bar and popover paths.
- Billing copy cleanup — Removed lifetime-specific checkout error copy in favor of generic active-plan wording.
- Offer set simplification — Activation and checkout now expose monthly/yearly offers only; lifetime stays supported only as a legacy entitlement path.
Fixed
- Settings popover clipping — Header no longer clips when Settings opens from the menu-bar gear icon.
- Settings freeze on startup — History restore no longer performs iCloud-backed transcription file reads on the main thread.
- iCloud placeholder safety — Startup restore skips non-materialized iCloud placeholders instead of blocking UI while files are downloading.
1.2.6 — February 6, 2026
Changed
- DMG packaging compatibility — Release packaging now uses an installer volume name to avoid a macOS copy restriction during DMG creation.
- Release channel sync — Updated Sparkle feed metadata and release documentation for build 14.
1.2.5 — February 6, 2026
Added
- Subscription plan offers in app — Activation now loads monthly and yearly options directly from pricing config.
Changed
- Plan-aware checkout — Checkout is now explicit per selected plan and currency, with stronger idempotency to prevent duplicate sessions on retries.
- Billing UX hardening — Activation and Settings now provide clearer recovery actions for pricing and billing-portal failures.
Fixed
- Post-pivot stability fixes — Addressed edge cases in grace-period handling, update-config timing, Sparkle integration, and checkout flow races.
1.2.4 — February 5, 2026
Added
- Onboarding wizard — 6-step guided setup for new users: permissions, microphone test, activation mode, and model download with progress indicator.
- Update notification in Settings when an update is downloaded and waiting for idle install.
Changed
- Safer auto-updates — Sparkle now defers update checks while recording, transcribing, or processing files; relaunch waits until Dicta has been idle for 20 minutes.
- Refined activation flow and permissions UX for a smoother first-run experience.
1.2.3 — February 5, 2026
Fixed
- Auto-detect language — Fixed issue where dictation in non-English languages was incorrectly transcribed to English when "Auto-detect" was selected.
1.2.2 — February 4, 2026
Added
- Auto-detect language option in Settings.
Changed
- Dicta now defaults to automatic spoken-language detection instead of system language.
- Live, file, and URL transcription all use the same auto-detect language pipeline.
1.2.1 — February 3, 2026
Added
- VoiceOver announcements for status changes (accessibility).
- Auto-update preferences in Settings.
- Help link in Activation view.
Changed
- Hotkeys enabled by default for new users.
- Improved activation UX and error messages.
- Refined setup banners and model retry UX.
- Dynamic popover height.
Fixed
- Handle Sparkle update check errors gracefully.
- Handle "purchase required" activation flow.
- ISO 8601 date parsing with milliseconds.
1.2.0 — January 31, 2026
Added
- Periodic license validation with a 24h retry window before deactivation.
- History browser window with split list/detail layout.
- Editable transcriptions with Cmd+S save and unsaved-changes prompts.
- Source metadata (URL/path) with Open/Copy actions in history detail.
- History actions: context menu, swipe-to-delete, Delete key.
Changed
- Activation now uses email sign-in codes (no link required).
- License transfers to another Mac when you sign in there (one active device at a time).
- History now opens in a dedicated window and closes the popover.
1.1.3 — January 31, 2026
Added
- Periodic license validation to prevent simultaneous device use.
Changed
- Signing in on another Mac transfers the license to that device.
1.1.2 — January 30, 2026
Added
- History recovery from transcription files (automatic restore if history.json is lost).
Changed
- Manual download now points to Cloudflare R2 instead of GitHub Releases.
1.1.1 — January 30, 2026
Added
- Help links for granting permissions in System Settings.
- Download fallback when binary downloads fail.
Changed
- Cleaner hotkey label: now shows "Opt + Space".
- Polished status bar and banner typography.
- Improved permission recovery after app updates.
- Better activation flow layout.
Fixed
- Swift 6 actor isolation warnings.
- Status bar popover layout stability.
1.1.0 — January 29, 2026
Added
- URL transcription: paste YouTube, Vimeo, Twitter/X links to transcribe audio (requires internet to download; transcription runs locally).
- FFmpeg audio extraction with broader format support (OGG, WMA, OPUS, MKV, AVI, WebM).
- Auto-download of ffmpeg and yt-dlp binaries (checks system PATH first).
- In-app onboarding for install and permissions.
- Permission checklist with deferred prompts.
- Clickable permission rows that open specific System Settings panes.
Changed
- Simplified main UI to two states: Error and Ready.
- History preview shows only last transcription (click to expand).
- Settings height now adapts to content.
Fixed
- YouTube 403 errors.
- WhisperKit cache path configuration.
- Timer memory leak in recording.
1.0.0 — January 18, 2026
Initial release
- Offline voice transcription with WhisperKit Large v3 Turbo.
- Privacy-first: audio never leaves your device.
- Global hotkey (Option+Space) for quick recording.
- ESC to stop and transcribe.
- Paste anywhere via accessibility.
- Transcription history with search.
- Launch at login option.
- Sound feedback toggle.
- Automatic updates via Sparkle.
- Microphone selection.
- Multi-language support (73 languages).
- iPhone microphone support via Continuity Camera.
Questions? Reach us at [email protected].
Like what you see?
Download free