Skip to content
ReleaseJuly 4, 2026Mallo Team2 min read

Mallo 1.2.24

Mallo 1.2.24 keeps the dictation HUD visible on the Mac desktop you are using and helps Qwen stay ready after idle time.

Mallo 1.2.24 makes dictation easier to trust when you work across multiple Mac desktops. The HUD now follows the desktop you are actually using, and Qwen is kept ready more gently after the app has been sitting open.

Release Highlights

  • The HUD follows your active desktop. If you switch between macOS Spaces, the dictation HUD should appear on the Space where you start dictating.
  • Fullscreen and multi-desktop work feels clearer. You should not have to jump back to another desktop just to confirm whether Mallo is listening.
  • Qwen stays ready after idle time. Mallo keeps the local Qwen path warmer by default so the next dictation is less likely to feel delayed.

What Changed In Real Use

If you use several desktops for coding, writing, browsing, or remote work, the old HUD behavior could feel confusing. Dictation might start, but the visual confirmation could stay on another Space.

This release makes the HUD reposition itself whenever dictation starts. It is a small fix, but it matters because the HUD is how you know Mallo heard the shortcut and is listening.

Qwen also gets a readiness pass. When Mallo has been idle for a while, the app now does more to keep the local model prepared without putting extra steps in your way.

Why This Patch Matters

Dictation should feel obvious in the moment: press the shortcut, see the HUD, speak, and keep working. This patch removes one more place where the app could feel detached from the desktop you were actually using.

Related updates

Keep reading

ImprovementMar 21, 20261 min read

Qwen and hotkeys are more ready when you are

We improved Qwen capture readiness and tightened hotkey behavior so dictation starts more reliably in the moments you actually reach for it.

Read update

Shipping continuously

Follow each release as Mallo gets sharper.

We’ll keep logging multilingual dictation, model setup, licensing, and workflow improvements here.