AI Auto Subtitle Creator

A PyQt6 desktop application for automatically generating and styling subtitles, built around OpenAI Whisper (Tiny through Large) for transcription accuracy, with a full dark-themed editing interface rather than a one-shot CLI script.

Pipeline

Load a video, pick a Whisper model on first run (downloaded once, then cached locally), transcribe. Output is word-level timed, which is what produces the karaoke-style highlight effect viewers expect from short-form video, rather than caption blocks that simply appear and disappear line by line.

Styling and editing

Built-in themes (Shadow 3D, Karaoke, Reel Bold, and others) cover the common short-form look out of the box, with full manual control over fonts, colors, outlines, shadows, and animations underneath, previewed live against the actual video rather than a static mockup. A timeline-based editor below the preview lets text and precise timing be adjusted per line before export.

Export

Three output paths depending on where the subtitles need to live next: .mp4 with subtitles burned directly into the frame via FFmpeg, or .srt / .ass subtitle files kept separate from the source video for platforms and editors that consume caption tracks natively. Transcription supports multiple languages, including English, Turkish, German, and French.