I am building a Windows-based desktop application that lets me load a golf-swing video—either a local MP4 or a YouTube link—mark eight key positions (from Address through Follow-through), and automatically turn those clicks into: • a precisely trimmed clip of the swing • an audio glide that spans each phase’s duration • synchronized exports in MP4, WAV, and JSON A detailed, feature-rich interface is essential: a scrub-able preview pane, frame-accurate in/out markers that update as I click, a timeline showing the eight swing points, and on-screen feedback of the generated tone. Under the hood the workflow I have in mind uses Python with FFmpeg for video clipping, Librosa or Tone.js for audio synthesis, and optional MediaPipe for landmark assistance; React (or a comparable framework) can drive the UI layer. The full scope covers: 1. Building the Windows installer and auto-update mechanism. 2. Video ingestion from file path or YouTube URL, with resolution-agnostic handling. 3. Interactive annotation: click-to-mark, keyboard nudge, and undo/redo. 4. Automatic audio generation: smooth frequency glide per phase, adjustable base tone. 5. Combined export that renders the annotated MP4, the standalone WAV, and a JSON file containing timestamps and metadata. 6. A preferences panel for tone parameters, default export folder, and branding elements. 7. Delivery of clean, well-documented source code plus a short screencast demo. I have a three-week window and an allocated budget in the $1.5k–$2.5k range. After selection I’ll share an NDA, sample swings, and my current UI sketches. When you reply, point me to a comparable video/audio processing tool you built, outline your approach, and confirm you can meet the timeline.