Vext v0.1,0

Vext v0.1.0 Initial Release

Vext is a local-first desktop speech-to-text utility focused on fast transcription, privacy, and seamless desktop text insertion workflows.

This initial release introduces a complete local transcription pipeline powered by whisper.cpp, along with a native desktop experience built using Tauri, React, TypeScript, and Rust.

Highlights

• Fully local speech-to-text transcription using whisper.cpp
• Hold-to-talk and toggle recording modes
• Clipboard-preserving text insertion on Windows
• Searchable local transcription history
• User-defined dictionary replacements
• Privacy-focused local storage model
• Cross-platform desktop foundation with Tauri

Core Features

Local Speech-to-Text

Vext captures microphone audio directly within the desktop application and performs transcription locally using whisper.cpp.

Key capabilities include:

• No cloud dependency required
• Support for multiple Whisper model sizes
• Browser-side WAV audio capture
• Native Rust-backed transcription execution

Desktop Text Insertion

After transcription completes, Vext restores focus to the previously targeted application and inserts text using a clipboard-preserving paste fallback on Windows.

Current implementation includes:

• Foreground window capture
• Clipboard restoration after insertion
• Compatibility-oriented insertion workflow

Dictionary Replacements

Vext supports customizable transcript cleanup through local dictionary replacements.

Included functionality:

• Phrase replacement rules
• Punctuation cleanup
• JSON dictionary export support

Local History and Statistics

All transcription history is stored locally on-device.

Included functionality:

• Searchable transcription history
• Aggregate usage statistics
• Export and deletion controls for local data

Desktop Interface

Vext includes a structured desktop UI with:

• Sidebar navigation
• Floating recording status bar
• Overview dashboard
• Settings management
• Onboarding and setup checklist

Platform Support

Windows

Windows is the primary supported platform for v0.1.0.

Requirements:

• WebView2 Runtime
• Microsoft Visual Studio Build Tools
• Enabled microphone permissions

macOS and Linux

Audio capture and preview functionality are available, but native insertion workflows are currently Windows-focused.

Testing

Basic test coverage is included for:

• Dictionary replacement behavior
• Statistics calculations
• Local storage management

Known Limitations

The following features are represented in the UI and settings model but are not fully connected to native OS hooks yet:

• Global hotkeys
• Always-on-top recording bar window
• Start-on-login behavior

Additional notes:

• Some elevated or sandboxed Windows applications may reject clipboard-based insertion
• Accessibility integrations for macOS are planned for future releases

Build Outputs

Expected Windows build artifacts:

src-tauri\target\release\Vext.exe
src-tauri\target\release\bundle\nsis\Vext_0.1.0_x64-setup.exe

Getting Started

Install dependencies:

npm.cmd install

Run development mode:

npm.cmd run tauri:dev

Run tests:

npm.cmd test

Build production executable:

npm.cmd run tauri:build

Closing Notes

Vext v0.1.0 establishes the foundation for a privacy-focused desktop dictation workflow with fully local transcription and native desktop integration. Future releases will expand native OS integrations, hotkey support, and broader platform compatibility.