tmoroney/auto-subs v3.5.1 on GitHub

Bug Fixes in v3.5.1

Custom text density: Added the ability to configure a specific maximum number of characters per line.
Korean formatting: Added a separate Korean script profile to preserve inter-word spaces
Cyrillic formatting: Fixed character handling, max chars-per-line, and traffic light button position
Resolve export: Corrected marker range calculation and fixed export cancellation handling
Audio track selection: Fixed scroll behaviour in the audio track selector
Model Manager: Recover gracefully from a broken Hugging Face cache and surface real download errors
Model Manager: Fixed delete button by bumping @tauri-apps/plugin-dialog to v2.7.0
Linux: Resolved install paths, render directory issues, and suppressed ORT log noise
Linux: Hid native titlebar to prevent double titlebar appearing
Subtitle list: Scoped CSS transitions to avoid WebKitGTK layer churn that was causing crashes on KDE

Voice Activity Detection — Automatically strips non-speech audio before transcription for cleaner, faster results
More Model Support — Added Whisper, Parakeet, and Moonshine to your transcription toolkit
Improved Model Management — Easier downloading, switching, and removal of models
Better Speaker Labeling — Upgraded diarization model for more accurate speaker identification
Smarter Formatting — Rule-based subtitle formatting optimised specifically for video
Translation Support — Translate your subtitles into other languages directly from the app

Freeform Subtitle Editing — The subtitle viewer now supports free-text editing (not just word-by-word), with automatic word timing adjustment
Transcript History — Browse and reload past transcripts without re-running the pipeline
Model Deletion — Clean up storage by removing models you no longer need
Custom Titlebar — A sleek new titlebar makes better use of app space and improves the overall look and feel
Localisation — The UI is now available in Korean, Spanish, German, French, Japanese, and Chinese 🌍

New Fusion Caption Macro — A custom macro that makes it simple to create animated captions with per-word highlighting
Preset System — Create your own caption presets, share them with others, or use one of the built-in options out of the box
Preset Preview — Preview how presets look directly inside Resolve before applying
Marker-Based Word Timing — Fine-tune word timing right on the Resolve timeline using simple markers
Conflict Detection — AutoSubs now checks for track conflicts the moment a track is selected

Exported audio is now saved to the app data directory instead of Downloads
Transcripts are also stored in the app data directory — a migration script automatically moves existing files from Documents
Better Error Reporting — Errors now pre-fill a GitHub issue template to make bug reporting effortless
General stability improvements across the board 🛠️

A big welcome to first-time contributors:

Thank you for helping make AutoSubs better! 💙

Full Changelog: v3.0.8...v3.5.1