embedded-captions

What problem does it solve?

Talking-head videos require professional, readable captions that integrate with the scene rather than obscure it, but manual motion graphics editing is time-consuming and demands specialized design skills.

Core Features & Use Cases

• 32 Visual Identities: Choose from cinematic column-flow styles (cream, ink, neon, glitch) or themed constitutions (anchor, ordnance, terminal, arcade) to match any tone from poetic to cyberpunk.
• Matte Occlusion Pipeline: Uses hyperframes remove-background to composite captions behind the subject, so the speaker's body naturally occludes text for a diegetic, embedded look.
• Deterministic Rendering: Transcribes audio via WhisperX, validates timing and occlusion gates, and composites via ffmpeg for reproducible outputs.
• Use Case: A YouTube educator records a 10-minute explainer and uses the 'anchor' identity to add clean verbatim lower-thirds with a single emphasized climax, or a music video director uses the 'neon' identity to make captions glow like signage behind the artist.

Quick Start

Use the embedded-captions skill to add cinematic captions to the video file 'monologue.mp4', choosing the 'cream' identity for a warm poetic look.