What’s new

The short version: Every release of TongueType, what changed, and when. Newest first.

1.0.10

September 9, 2026

  • TongueType now learns placeholder text in apps that present it as real content, like the Claude Code extension’s message box in VS Code. When a dictation replaces such text, TongueType remembers it as a placeholder, so from then on dictating over it starts your sentence clean.

1.0.9

September 9, 2026

  • Smart insertion now fixes casing in both directions: it capitalizes the first word when you start a new sentence, at the beginning of a field, after a new line, or after sentence-ending punctuation, and still lowercases it when you’re continuing one.
  • Fixed a bug where dictating after a sentence in web-based editors like Spark could start with a lowercase letter, caused by the invisible non-breaking space these editors place before the cursor.
  • Dictating into a terminal no longer starts with a lowercase letter: when TongueType can’t tell what the text before the cursor is (like a shell prompt), it now leaves your transcription’s casing alone.
  • Sentences that end in a closing quote or bracket, like He said “Stop!”, are now recognized as sentence ends.
  • Fixed the first dictation into an empty email in Spark coming out lowercase with a stray space in front. Spark’s compose field shows placeholder text that looks like real content to macOS accessibility; TongueType now recognizes placeholders and treats the field as empty.
  • Fixed back-to-back dictations running together (“Test one.Test two.”) in code editors like VS Code and Cursor. These editors hide their text from macOS accessibility, so TongueType can’t see where your cursor sits; when it detects such an editor, it now ends each transcription with a space so the next one stays separated.

1.0.8

August 7, 2026

  • TongueType now adapts transcriptions to the text around your cursor: it starts with a lowercase letter when you’re continuing a sentence, like when you resume dictating after a break, and adds a missing space when your cursor sits right against a word or punctuation. It works in most text fields, but some apps, like certain terminals and code editors, don’t expose their text, in which case transcriptions are inserted unchanged. A new Smart insertion toggle in Settings turns it off if you prefer.
  • Caps Lock now works with dictation: leave it on and your transcription comes out in ALL CAPS.
  • Merged the “Accent color” and “Theme” controls in Settings → Appearance into a single section.
  • Reworded the post-processing UI to make it clear that you can replace matches with symbols, words, or phrases.

1.0.7

May 29, 2026

  • The hotkey grace period can once again be set all the way down to 0 ms, and 0 ms is the default. A change in 1.0.6 had accidentally raised the minimum to 100 ms.

1.0.6

May 22, 2026

  • Pro licensing now lives in its own tab in Settings instead of a separate window. The menu’s  Unlock Pro… item opens straight to it.
  • The hotkey grace period now ranges from 100 ms to 1000 ms, with 100 ms as the default.
  • The recording overlay now stays on screen after you release the hotkey, showing a progress bar while your words are transcribed.
  • Fixed a bug where brief recordings sometimes came back empty.
  • Fixed double-tap-to-latch which was unreliable whenever the hotkey grace period was non-zero.
  • Translated previously-untranslated terms (license and Pro screens, setup wizard warnings, menu warning rows)
  • You can now press Return in a postprocessing rule to add a new rule below it, so you can build up your rules without reaching for the mouse.
  • Added color themes for the recording overlay — six multi-color palettes that animate the indicator and waveform, alongside the existing Rainbow mode.
  • Added a Zinc accent color.

1.0.5

May 14, 2026

  • Fixed a bug that could leave the loading indicator stuck on screen after launch.
  • Fixed the “Lost your key?” link so it opens Polar instead of an email.

1.0.4

May 12, 2026

Initial release.

  • Hold-to-record voice dictation with a configurable hotkey.
  • Optional double-tap-to-latch for hands-free dictation.
  • 11 languages plus auto-detect.
  • Audio and video file transcription via drag-and-drop or file picker.
  • Customizable postprocessing rules ("new line", "open parenthesis", and so on).
  • Recording overlay with multiple anchor positions, color schemes, and accent colors.
  • Microphone selection with automatic fallback and reconnect.
  • Output mode: insert at cursor (default) or copy to clipboard.
  • First-launch setup wizard for model download and permissions.
  • 100% on-device processing via WhisperKit on Apple’s Neural Engine. No accounts, no servers, no analytics.
  • Built-in automatic updates via Sparkle.