See it in action

A real render from the kit: word-synced captions, real b-roll, an animated overlay, and a thumbnail. No watermark.

Each scene is one short object: a template, a headline, and the voiceover line. Write it by hand, or have an AI write it from a prompt.

Prompt (the scene schema is in the kit docs):

Write a 45-second faceless YouTube Short as the kit's scene JSON:
a hook, two focus tips, and a CTA to the Faceless YouTube Kit.
Keep each voiceover line under 20 words.

It returns a scene file like this:

{
  "slug": "focus-showcase", "format": "portrait",
  "scenes": [
    {"template": "hook", "h1": "Can't focus at home?", "bg": ["typing-laptop.mp4"]},
    {"template": "tip", "badge": "1", "h1": "One screen rule", "lines": ["Close every tab but one"]},
    {"template": "cta", "h1": "Which will you try?", "card_title": "Faceless YouTube Kit"}
  ]
}

The kit renders the video above, plus three vertical Shorts and a thumbnail, from that one file.

The wiring nobody wants to debug twice.

Getting edge-tts word timings into ffmpeg ASS captions, deriving three Shorts from a single long-form render, and fetching b-roll without burning an API budget takes days to wire up correctly. This kit has it working, documented, and tested.

What you get

Long-form + three Shorts from one file

Write one scene JSON, run one command, and get a 16:9 deep-dive with chapters and up to three vertical 9:16 Shorts ready to upload.

Word-synced karaoke captions

edge-tts word timings feed directly into an ffmpeg ASS subtitle burn, so every word highlights exactly when it is spoken.

Thumbnail generator included

make_thumbnail.py renders a branded still from your scene JSON using the same headless-browser pipeline as the video.

B-roll fetcher for Pexels and Pixabay

fetch_pexels.py and fetch_pixabay.py pull royalty-free clips to your local assets/ folder; the renderer picks them up automatically.

Guided first-run setup

setup.py walks through API keys, fetches starter CC-BY music with attribution, and runs a smoke-test render so you know everything is wired before you write a scene.

Full scene-JSON schema in docs/

GUIDE.md documents every field: hook, tip, dialogue, cta, music bed, broll, Ken Burns, chapter markers, and the retention method behind the scene order.

From the writeup, packaged · just €29

Skip the rabbit hole. Ship it.

The exact, working implementation, generalised so you drop it into your own project and move on.

Get the kit — €29

Instant download · single-project commercial use · ✓ 30-day money-back guarantee