ensu-v0.1.16

thmoa/tex2shapeensu-v0.1.16May 25, 2026by github-actions[bot]

AI Summary

This release introduces significant enhancements to the desktop and mobile applications, focusing on AI model integration, media handling, and download performance.

Key Highlights

Added Gemma 4 model support for the desktop app.
Introduced local voice transcription for chat input on iOS and Android.
Improved image attachments with compression, full-screen previews, and cleaner composer previews.
Optimized model downloads with parallel/range downloads, retries, and better progress reporting.

New Features

Local voice transcription for chat input
Image compression and full-screen previews
In-app changelogs for desktop, Android, and iOS
Improved desktop model selection with better system memory detection

Full Release Notes

- Added Gemma 4 for the desktop app
- Added local voice transcription for chat input on iOS and Android.
- Improved image attachments with compression, thumbnails, full-screen previews, and cleaner composer previews.
- Improved image prompt performance on native apps by resizing to model limits and caching multimodal context.
- Add in-app changelogs to desktop, android, and iOS
- Balanced haptics on mobile to once at the start of response generation
- Improved model downloads with shared native downloading, cache reuse, parallel/range downloads, retries, and better progress reporting.
- Improved desktop model selection by using better system memory detection.
- Fixed Android back button handling when the sidebar is open.
- Fixed the desktop app version shown in settings.