Whisper Desktop

Whisper Desktop turns audio and video recordings into transcripts and subtitles on your computer. It's for people working with interviews, lectures, podcasts or videos who want a desktop interface for transcription and captioning.

• Transcription and batches: Process individual recordings or a queue of files, then export text and subtitles in formats including SRT, VTT, TXT, JSON, ASS, LRC and CSV.
• Subtitle translation: Translate existing subtitle files through a chosen AI provider or local LLM while keeping their timings intact.
• Permanent captions: Burn subtitles into a video with adjustable styling and GPU-accelerated FFmpeg rendering.
• Model selection: Browse Whisper GGML models, including English-only and quantized variants, with hardware-based recommendations and memory estimates; the model manager downloads and verifies the files.

The app runs on Windows, macOS and Linux. It bundles whisper.cpp and FFmpeg, so transcription doesn't require a separate model server or cloud account. CPU-only inference works, with optional NVIDIA, AMD, Intel or Apple Silicon acceleration through CUDA, Vulkan, OpenVINO or Metal, depending on the platform and package.

Models aren't bundled, and downloading them requires internet access. Once models are available locally, transcription and subtitle burn-in work offline. Translation uses the selected provider; choosing a cloud service sends subtitle text outside your computer.