transvibe

Transcribe+Translate: Real-time, local transcription and translation.

Note

This project is a "vibe coding" session with Gemini 2.5 Pro Preview, developed using the Zed Editor.

Features

• 🦀 Built with Rust: Crafted by Rustaceans using Kalosm for AI and Ratatui for the terminal interface.
• 🏡 100% Offline: Operates entirely locally, ensuring your data privacy and functionality without an internet connection.
• 🎤 Real-time Transcription: Captures and transcribes audio from your microphone as you speak.
• ⚡ Responsive Translation: Utilizes a separate thread for translation, preventing delays in the transcription process.
• ✨ Enhanced Readability: Highlights the most recent transcribed line, making it easier to follow along.

TODO

Future enhancements, planned features, and currently missing capabilities include:

• Expanded Language Support: Adding transcription and translation capabilities for languages beyond the current scope.
• File-based Transcription: Implementing the ability to transcribe audio from various file formats (currently missing).
• Saving and Exporting: Adding functionality to save and export transcriptions and translations (currently missing).
• User-configurable Settings: Implementing options like audio input device selection (currently missing).
• Pause/Resume Functionality: Adding the ability to pause and resume transcription (currently missing).
• Session Management: Allowing users to save and load transcription/translation sessions.
• UI/UX Improvements: Continuously refining the user interface and experience.
• Performance Optimizations: Further optimizing processing for speed and resource efficiency.
• Customizable Models: Allowing users to specify different speech-to-text or translation models.

Run

To run the application:

cargo run --release

Build

To build the application from source:

cargo build --release