Active development · media accessibility
Aura Engine
A long-form video pipeline designed to make foreign-language media easier to follow through English dubbing, queue-based processing, and synchronized playback.
The problem
Long-form foreign-language video is difficult to consume when reading subtitles is uncomfortable or impractical.
The project focuses on an audio-first experience: locate a playable stream, prepare media, coordinate subtitles and speech, then keep dubbed audio synchronized with the original video.
Current architecture
A React and Vite interface coordinates a Node sidecar that discovers playable MP4, HLS, or DASH streams.
The interface manages a playlist queue, extraction progress, playback state, and the future processing handoff for subtitle recognition, translation, and speech synthesis.
Engineering work so far
The most important work has been state integrity rather than surface styling.
I have been tracing deletion, reordering, cancellation, stale async updates, manual selection, and extraction progress so that the visible queue matches the actual work being performed.
Why it matters
Aura Engine demonstrates product thinking across a full media workflow.
The project combines browser playback, backend extraction, asynchronous job state, accessibility goals, and future AI speech integration rather than treating dubbing as a single API call.