What problem does it solve?
When adding captions to talking-head or launch videos, creators often misplace content by reserving a dead bottom band for subtitles or overusing embedded word effects, producing unbalanced compositions. This Skill defines the caption model and overlay rules so captions composite cleanly on top of the film without distorting the layout.
Core Features & Use Cases
- Caption Model (drop / rail / embed): Classifies every spoken phrase as dropped filler, verbatim lower-third rail text, or a scarce embedded peak word composited behind the subject.
- Overlay Law Enforcement: Ensures captions are treated as an overlay layer on the true vertical center of the frame, never a reserved keep-out band that shifts content upward.
- Use Case: When building a product launch video with word-by-word captions, use this Skill to center the composition at y = H/2, keep the verbatim rail readable, and promote at most one word per beat to an embedded climax.
Quick Start
Use the captions-overlay skill to add rail-first captions with one embedded peak word to my talking-head video composition.
