vawe-type-talking-head

Plans narrated talking-head videos with word-timed captions, face-safe layouts, and B-roll cut rhythm.

3|Updated Jun 9, 2026
One-click install
npx skills add https://github.com/Vandit1604/vawe --skill vawe-type-talking-head-vandit1604
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vawe-type-talking-head
Source: https://github.com/Vandit1604/vawe/tree/main/skills/vawe-type-talking-head
Command: npx skills add https://github.com/Vandit1604/vawe --skill vawe-type-talking-head-vandit1604

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Authoring a narrated presenter-style video in the Vawe engine requires coordinating voiceover timing, captions, face-safe composition, and B-roll cuts, which generic scene templates do not handle. ## Core Features & Use Cases - VO-driven pacing: Builds beat timings from a real voiceover word-timing sidecar using make pace-from-vo and generates karaoke-timed captions with make vo-captions. - Face-safe composition: Enforces a quiet backdrop register, a clear safe centre for the presenter, a persistent lower-third, and caption-safe strips for phone destinations. - B-roll cut rhythm: Structures the film as presenter to proof cutaway to presenter, using devices like statReveal and containerFill for evidence beats. - Use Case: Given a 12-second 9:16 shorts brief with a recorded voiceover, scaffold a talking-head scene, derive beat timings from the VO words file, and produce captions that land exactly on the spoken words. ## Quick Start Ask the agent to plan a narrated talking-head video from your voiceover file and brief using the talking-head playbook.

Frequently Asked Questions about vawe-type-talking-head

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a talking-head video with word-timed captions?

Provide a voiceover track plus its word-timing sidecar, then run `make vo-captions` to build karaoke-timed captions and `make pace-from-vo` to align beat reveals with the spoken words. Scaffold the scene with `make scaffold TYPE=talking-head`.

What layout rules apply to presenter-style videos?

Keep the safe centre of the frame free for the face, hold the backdrop dark and slow in the quiet register, and place name and title in a low fixed lower-third. On phone destinations, also respect the caption-safe strip for platform chrome.

Can I make a talking-head video without real presenter footage?

Yes, but the presenter slot must be a clearly labelled placeholder rect declared in the scene note, never invented UI. The same applies to VO: use synthesized speech from `make tts` as a draft until a real recording exists.

When should I not use the talking-head playbook?

Avoid it for videos driven by a visual device or transforming object rather than a voice; those belong to the continuous-action or other routed types. Talking-head pacing is set by the voice track, with the slowest beat band of 3.0-5.0 seconds.

Why do my captions not match the voiceover timing?

Hand-timed captions drift from real speech. Generate them from the VO word-timing sidecar with `make vo-captions` instead of guessing durations, so each caption lands on the actual spoken words.