Wiss Shorts
US$ 10
The full skill: script, voice, acting characters, 46 mascots, 92 styles, 4 sizes, cover and subtitles, in English or Spanish. You ask for changes in the chat.
Buy without editorA skill for Claude Code, Codex or your favourite agent
Give it a topic and get the video back: script, voice, characters that act, music and subtitles. Vertical, square or horizontal.
Pick a topic and see what Wiss Shorts does with it: it writes the script, splits it into shots and delivers the video. These are the real videos above; the request is summarised.
Loading the engine…
If you'd rather tweak it yourself, the short opens in a visual editor in the browser, on your computer. Change the text, the style, the music or a mascot and see it instantly, without remaking the video.
Comes with Wiss Shorts + Editor (US$ 15), or add it later for US$ 5.
Every short comes with 12 covers to choose from. The character cuts itself out and the headline moves around, or the video shows up on a TV, a newspaper, a magazine, a laptop, a billboard or a phone. Always inside the part the TikTok and Instagram grid shows, and horizontal too, as a YouTube thumbnail.
A whole video without a single cut: the camera travels through one pixel scene, zooming in or out from one thing to the next. It works for any topic: a story, a product, a fun fact, a challenge or a trip like this one, from you to the galaxy and back.
hola. It installs itself and makes a test short./wiss-shorts or just: “make a short about why flamingos are pink”.Free mini tutorial: how to make your own video skill with Claude Code. 5 lessons with real code (about 45 minutes). At the end you'll have a simple short: a headline that comes in word by word on the beat of a voice made on your computer.
Claude Code reads SKILL.md: the description tells it when to use the skill, and the steps tell it what to do. The heavy lifting (voice, timings, render) lives in scripts Claude runs for you. Copy it to ~/.claude/skills/my-short/.
---
name: my-short
description: Makes a narrated vertical short. Use it when asked to "make a short about X".
---
1. Write a 6–8 short sentence script and fact-check every claim.
2. Make the voice: python tools/voz.py "<script>"
3. Measure the timings: python tools/tiempos.py
4. Paste tiempos.json into index.html and render: npx hyperframes render --output short.mp4
5. Make the contact sheet and check it before delivering.Kokoro-82M speaks English (and Spanish), it's free and needs no account: your script never leaves your machine. Install kokoro-onnx and soundfile with pip and download the two model files (kokoro-v1.0.onnx and voices-v1.0.bin).
import sys
import soundfile as sf
from kokoro_onnx import Kokoro
k = Kokoro("kokoro-v1.0.onnx", "voices-v1.0.bin")
audio, sr = k.create(sys.argv[1], voice="am_michael", speed=1.0, lang="en-us")
sf.write("voz.wav", audio, sr)To bring the text in on the beat of the voice you need the moment of every word. faster-whisper measures it locally (pip install faster-whisper numpy). If a word comes out wrong, change "base" to "small".
import json
import numpy as np, soundfile as sf
from faster_whisper import WhisperModel
# whisper wants mono audio at 16 kHz
y, sr = sf.read("voz.wav")
y = y.mean(axis=1) if y.ndim > 1 else y
a16 = np.interp(np.arange(0, len(y) / sr, 1 / 16000), np.arange(len(y)) / sr, y).astype(np.float32)
m = WhisperModel("base")
segs, _ = m.transcribe(a16, word_timestamps=True, language="en")
palabras = [[round(w.start, 2), w.word.strip()] for s in segs for w in s.words]
json.dump(palabras, open("tiempos.json", "w"), ensure_ascii=False)HyperFrames turns an HTML page into an MP4, frame by frame. GSAP animates with a paused timeline: every frame comes from the time, so the result is always the same. Here, each word pops in right when it's spoken.
<body>
<div id="root" data-composition-id="main" data-start="0" data-width="1080" data-height="1920" data-duration="3">
<!-- without an id the audio is silent; data-duration = length of voz.wav -->
<audio id="voz" src="voz.wav" data-start="0" data-duration="2.4" data-track-index="1"></audio>
<h1 id="titular" style="font: 700 96px sans-serif; margin: 300px 80px"></h1>
</div>
<script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>
<script>
const palabras = [[0, "Did"], [0.2, "you"], [0.36, "know"], [0.6, "flamingos"], [1.2, "are"], [1.4, "born"], [1.7, "grey?"]]; // tiempos.json
const tl = gsap.timeline({ paused: true }), h1 = document.getElementById("titular");
for (const [t, w] of palabras) {
const s = document.createElement("span"); s.textContent = w; h1.append(s);
Object.assign(s.style, { display: "inline-block", marginRight: "0.25em" });
tl.from(s, { y: 40, opacity: 0, duration: 0.3, ease: "back.out(2)" }, t);
}
window.__timelines = { main: tl };
</script>
</body>A sheet with two frames per second to see it all at a glance, and loudness at −14 LUFS, what TikTok, Reels and Shorts expect.
ffmpeg -y -i short.mp4 -vf "fps=2,scale=180:-1,tile=10x6" hoja.jpg
ffmpeg -y -i short.mp4 -c:v copy -af loudnorm=I=-14:TP=-1 final.mp4With the above you have text that comes in on the beat of a voice. For a short to hold attention you need all of this, and it's the part that takes weeks of trial and error:
One-time payment, no subscription and no credits: everything runs on your computer. The videos you make are yours, to publish and monetise.
US$ 10
The full skill: script, voice, acting characters, 46 mascots, 92 styles, 4 sizes, cover and subtitles, in English or Spanish. You ask for changes in the chat.
Buy without editorUS$ 15
Everything above plus the visual editor: preview, shots, styles, draggable elements, sound, cover and render, without touching code.
Buy with editorAlready have Wiss Shorts without the editor? Add the editor for US$ 5: it installs on top with one command.