Changelog

What’s new
in Kudu.

Every release, newest first. The full notes and downloads are on GitHub.

  1. v0.1.2October 15, 2026Latest

    More ways to install

    • A Windows installer: per-user, in the Start menu, and removable from Settings → Apps.
    • Linux packages for Debian and Ubuntu (.deb) and Fedora and openSUSE (.rpm), which set up the virtual keyboard for you.
  2. v0.1.1October 15, 2026

    Downloads fixed on Windows and Linux

    • The app no longer closes while downloading the speech model.
    • A failed download or dictation shows an error and Kudu carries on.
    • Crashes are written to kudu.log, so they can be diagnosed.
  3. v0.1.0October 14, 2026

    First public release

    • Kudu for macOS (Apple Silicon), and Kudu for Windows and Linux in beta.
    • One-command install scripts that also clear the unsigned-app warnings.
    • Customizable shortcuts for dictation, Command mode and paste last transcript.
    • On Windows and Linux: Parakeet or Moonshine speech, and optional local formatting.

Before 0.1

The development milestones that led to the first release.

  1. v0.7Milestone

    Spelling that sounds right

    • Sound-alike matching against your dictionary and the names on screen.
    • Spoken file names: “use case dot ts” → useCase.ts.
    • “Added to dictionary” in the pill, with Undo.
  2. v0.6Milestone

    Formats like a native writer

    • Lists, units, times and math notation.
    • Code editors and terminals get symbols and identifiers.
    • Names spelled the way they appear on screen.
  3. v0.5Milestone

    Learning and Insights

    • Learns from your corrections, on-device.
    • Insights: words per minute, time saved and a streak heat map.
  4. v0.2–0.4Milestones

    The foundations

    • Styles per app, snippets and the dictionary.
    • Command mode for rewriting selected text.
    • The menu-bar app and the recording pill.

Looking for a specific build? All releases on GitHub →