mehtabmahir/easy-whisper-ui

What it solves

EasyWhisperUI provides a user-friendly graphical interface for the Whisper model, allowing users to transcribe audio and video files locally on their own hardware without needing to rely on cloud services or complex command-line tools.

How it works

The application is built with Electron and leverages whisper.cpp and FFmpeg to perform transcription. It supports GPU acceleration via Vulkan (Windows) and Metal (macOS) to speed up processing, while also providing a CPU-only fallback. Users can choose from various Whisper models, manage custom models, and export results as TXT or SRT files.

Who it’s for

It is designed for people who want private, offline transcription of audio and video, including those who whom need to generate subtitles (SRT) or prepare text for further processing with LLMs.

Highlights

• Local and Private: Transcription happens entirely on the user's machine.
• GPU Acceleration: Automatic verification and support for Vulkan and Metal.
• Batch Processing: Ability to drag and drop multiple files into a queue.
• Multilingual Support: Supports transcription and translation into English.
• Live Transcription: Beta support for real-time microphone audio transcription.
• Model Management: Easy downloading and loading of custom models with memory estimates.