Uses the Whisper models to transcribe videos into subtitles (Rust and Slint) (github.com)

A new cross‑platform desktop app leverages OpenAI’s Whisper models to transcribe audio and video into editable subtitles, with built‑in AI translation and AI‑assisted correction. The app (implemented in Rust with a Slint UI) supports Linux, Windows and macOS, lets you download Whisper models locally, play media while correcting transcripts, edit subtitles, and export subtitle files or burned‑in video. It’s aimed at streamlining subtitle workflows — from raw ASR to translation and final export — while keeping processing on the desktop for privacy and low-latency work. Key technical notes: the project is built in Rust (requires cargo) and provides make targets (make desktop-debug and make desktop-build-release). It uses ffmpeg for audio/video format handling (ffmpeg must be on PATH on Windows), and Linux builds require Zenity or KDialog for file dialogs. The Qt backend is recommended — especially on Windows — to avoid fuzzy fonts and to match developer build environments. By packaging Whisper locally and combining playback, editing and AI correction/translation, the app demonstrates a practical, production‑oriented desktop deployment of ASR+MT pipelines suitable for content creators and researchers who need offline, editable subtitle workflows.