V5: Data Intelligence, Voice, and Karaoke
Data intelligence. Voice synthesis. Karaoke subtitles. And a workspace that feels like VSCode. V5 is here.
Five weeks ago, we shipped V1. Now we're at V5—and each release has been bigger than the last. This one? It turns Diwadi from a file manager with AI into something much more powerful: a data intelligence platform with voice capabilities.
What's New in V5
Data Intelligence Suite
Create interactive dashboards from your CSV and Excel files. Add calculated fields with custom formulas. Build charts, comparison visualizations, and resizable widgets. Save everything as .dwb workbook files. It's like having a mini Tableau built into your file manager.
Voice & Audio System
Text-to-speech that works completely offline. Speech-to-text transcription built in. No internet required, no API keys needed. Your voice data stays on your machine.
Karaoke Subtitles
Generate word-by-word timed subtitles from any video. Automatic transcription creates precise word-level timing. Burn subtitles directly into videos. Perfect for music videos, tutorials, or social content.
Multi-Tab Workspace
VSCode-style split panes. Open multiple files side-by-side. Tab management with synchronized state. Work the way developers work—but for all your files.
Enhanced Video Editor
Audio editing with magnetic alignment. Text overlays with duration control. Image drops to timeline. Full undo/redo history. Save projects as .vep files with all assets bundled.
10 Complete Languages
10,451 translation keys across 10 languages. Pre-commit hooks ensure no hardcoded strings slip through. Missing key detection keeps translations complete.
The Technical Details
Complete refactoring of video/subtitle/caption logic. Header and file navigation reorganization. PDF parser improvements. Layout and tabstore restructuring. Dead code cleanup throughout the codebase.
The Numbers
156 commits since V4. 30+ bug fixes. Complete data intelligence suite. Offline voice system. All still fitting in a lightweight download.
Five Weeks of Shipping
V1 on Oct 31. V2 on Nov 7. V3 on Nov 14. V4 on Nov 21. V5 today. We're building in public and shipping every week. This is how you build software that people actually want.
View Full Changelog →Five releases. Five weeks. Download V5 now and see how far we've come.
Download Diwadi V5Full Changelog
Major Features
- Data Intelligence Suite with dashboards, calculated fields, and visualizations
- Widget system with resizable, formattable widgets and direct data binding
- .dwb workbook format for saving and loading dashboard projects
- Offline text-to-speech—no internet required
- Speech-to-text transcription built in
- Karaoke subtitle generation with word-level timing
- Burn subtitles directly into videos
- VSCode-style multi-tab workspace with split panes
- Audio editor with magnetic alignment to video clips
- Text overlays with duration control and design options
- AI-powered Word document creation
- Complete undo/redo history for video editing
- .vep video project format with bundled assets
Technical Improvements
- 10 complete language translations with 10,451 keys
- Pre-commit hooks for hardcoded string detection
- Intelligent git staging that re-stages only previously staged files
- Bun-only testing framework enforcement
- Optional native dependencies for headless CI builds
- Video/subtitle/caption logic extraction and refactoring
- Header and file navigation reorganization
- PDF parser improvements
- Layout and tabstore restructuring
UI/UX Enhancements
- Activity panel redesign with better cards
- File browser search within browser panel
- Theme-based logos on splash screen
- File tree tooltips for full filenames
- Improved video preview loading states
- Better loader animations throughout the app
- Current word highlighting during karaoke playback
- Auto-scroll to generated captions