easy-whisper-ui: Transcribe Audio and Video Locally

Overview

EasyWhisperUI provides a graphical way to run Whisper transcription on your own computer. It uses whisper.cpp to process audio and video locally, helping people who want transcription without uploading media to an online service or managing the model from a terminal.

It is aimed at desktop users who want a straightforward workflow across Windows, macOS, and Linux. GPU acceleration is available on supported hardware, while CPU-only processing is also possible.

Key Features

• Transcribes audio and video, with a beta live-transcription option.
• Queues multiple files for sequential batch processing.
• Exports plain text (.txt) and timestamped subtitles (.srt).
• Supports translation into more than 100 languages.
• Converts media to the required format using FFmpeg.
• Offers model and language selection, automatic model downloads, and custom local model files.
• Supports Vulkan acceleration on Windows and Metal on Apple Silicon Macs.

Use Cases

• Content creators can turn recordings or video files into text and subtitle files.
• Researchers and students can transcribe interviews or recorded sessions locally.
• People handling sensitive recordings can avoid sending media to a transcription service.
• Users with supported GPUs can process recordings through a desktop interface with hardware acceleration.
• Users who need multiple recordings transcribed can queue files instead of processing them one at a time manually.

Getting Started

Download an installer for your platform from GitHub Releases, then follow the platform's installation steps. See the README for requirements and setup details.

Alternatives

• voicebox: Voicebox focuses on dictation, speech generation, and voice cloning, while EasyWhisperUI focuses on batch and live transcription with Whisper.cpp.

Considerations

• GPU acceleration depends on compatible hardware: Vulkan on Windows and Apple Silicon with Metal on macOS. CPU-only operation is supported but described as slow.
• The README identifies live transcription as beta, and Linux support is available with a request to report distro or hardware issues.
• The application is described as proprietary: commercial use, copying, modification, and distribution require the author's permission. Personal use is permitted under the stated terms.
• On first Linux launch, the app installs required dependencies and compiles Whisper binaries.