Whisper Desktop v2.3.0
🚀 Whisper Desktop v2.3.0
v2.3.0 is officially live! A major release featuring complete UI/UX modernizations, hybrid TypeScript architecture, real-time downloader telemetry, and an upgraded AI translation engine.
🌟 What's New in This Release?
🎛️ 1. UI Modernization & Top-Layer Portal Dropdowns
- Top-Layer Dropdown Menus: Select menus now render seamlessly in the top layer, completely eliminating dropdown clipping across all overflow boundaries and scroll containers.
- Unified Batch Processing Queue: The multi-file batch queue has been redesigned into a cohesive glassmorphic panel featuring an integrated header bar, live file count badges, and aligned action controls.
- Refined Visual Affordance: Improved card hover behavior across all views—eliminating white flashes on static layout containers while preserving soft accent highlights on interactive controls.
- Interactive Toast Notifications: Enhanced system notifications with animated progress countdowns, hover-to-pause interactions, and quick close controls.
📥 2. Real-Time Model Downloader & Telemetry
- Event-Driven Progress Streaming: Download progress updates dynamically in real time without refreshing the page or interrupting navigation across other tabs.
- Standardized Mbps Telemetry: Download speeds are now calculated and displayed in standard Megabits per second (Mbps).
- Smart ETA Calculations: Estimated completion times exceeding 60 minutes automatically format into clean hours, minutes, and seconds.
- Smoother Speed Accuracy: Baseline file tracking eliminates initial speed spikes when resuming partial downloads.
- Clean Pause & Resume: Pausing an active download is tracked gracefully without generating false error notifications.
🌐 3. Multi-Language AI Subtitle Translation Suite
- 40+ World Languages Support: Expanded translation support for over 40 global languages.
- Sliding Context Window: Keeps conversational context intact across long media subtitle segments to ensure natural, narrative dialogue.
- High-Fidelity Dialogue Prompting: Subtitle translations preserve speaker formatting, punctuation, line numbers, and exact pronoun fidelity.
- Real-Time Chunk Progress: Stream line-by-line translation progress directly inside the interface as AI processing occurs.
⚡ 4. Modern TypeScript & Vite Toolchain
- Hybrid TypeScript Support: The frontend build system now supports TypeScript, enabling modern, strongly typed module development alongside existing JavaScript components.
- Blazing Fast Build Performance: Integrated Vite tooling for near-instant development updates and optimized production builds.
🛠️ 5. Engine Reliability & Diagnostics
- Stabilized Output Monitoring: Improved stdout reader stability during active transcription sessions.
- Automatic Silero VAD Synchronization: Automated state alignment between Voice Activity Detection settings and command-line execution parameters.
📜 Full Commit Log
be78b13chore(release): bump version to v2.3.04447bfbfeat(infra): configure typescript and vite hybrid environment for future module development5d317e7feat(dashboard & downloader): update dashboard feature highlights, refactor model downloader events, and fix pause/speed telemetrye115982fix(ui): refactor batch queue card layout and implement top-layer portal positioning for custom selectseec30bestyle(ui): eliminate noisy white card hovers and standardize affordance across components6ed6eb4style(ui): redesign models storage directory field into responsive read-only path control14b4b57feat(translation): expand target languages list to support 40+ major languagesb092a70fix(translation): balance system prompt for complete fidelity and natural spoken dialogued23433afeat(translation): emit per-chunk progress events and cap chunk size at 30 items301e34cfix(transcribe): prevent premature stdout reader loop termination and fix misleading abort notification277a77efeat(translation): implement sliding context window and conversational subtitle prompt optimizationc1b44ddfeat(ui): upgrade toast notifications with longer durations, progress bar countdown, hover-pause, and close button
📥 Getting Started
Download the package corresponding to your Linux distribution from the Assets section below:
- Ubuntu/Debian:
.deb - Fedora/RedHat:
.rpm - Portable Linux:
.AppImage
Arch Linux (AUR):
Arch Linux users can install the application directly from the AUR:
paru -S whisper-desktop-bin(or using your preferred AUR helper like yay)
Enjoy high-performance, private, and offline speech-to-text transcription! 🎙️✨
Full Changelog: v2.2.0...v2.3.0