Faceless Shorts Factory

What you're buying: a download package (skill instructions + Python scripts) that runs on your own computer inside your own AI agent — it is not a cloud service.
You need: an AI agent that supports skills (e.g. Claude Code or Codex) — its subscription or API costs are yours; Python 3.9+ and ffmpeg.
Third-party services: Pollinations.ai images (free tier: 1 image per 45 s, small watermark; an optional paid key removes the limit and watermark) and Microsoft edge-tts voices (free, internet needed).

One topic in, one ready-to-upload vertical short out. No face on camera, no editing software.

What you get

For every video:

• 🎬 1080×1920 MP4: AI-generated visuals, neural voice-over, burned-in subtitles, push/pull camera motion and background music
• 📝 SRT subtitle file
• 🖼️ 1280×720 thumbnail: big headline with the key word in an accent colour
• 📋 Upload copy: 5 title options (best one marked ⭐), description, hashtags and a pinned comment

Example

Topic: "How big is the Sun, really?" → a 20-second English short with 5 scenes, voice-over, subtitles and motion, plus the thumbnail "THE SUN IS HUGE — but how huge?" and upload copy.
A 7-scene Traditional Chinese science short (46 s) was made the same way.

How it works

1. Give a topic — or let your agent pitch 5 ideas from 10 proven viral formulas.
2. The agent writes the script with a built-in hook → build-up → payoff → CTA structure and asks you to approve it.
3. The script generates images (Pollinations), voice (edge-tts), subtitles and the final video.
4. The agent checks a frame for subtitle legibility, then hands over the files and upload copy.

Batch mode: "Make 5 videos about ___" → it lists the topics first, then renders them one by one.

Languages

400+ edge-tts voices, including English, Traditional/Simplified Chinese, Japanese and Korean. Subtitles wrap by visual width and never split English words or numbers.

Requirements

• Python 3.9+ and ffmpeg (the skill checks your setup and tells you what to install)
• pip install -U edge-tts pillow
• Internet connection

Speed and free tier

• The free image tier allows 1 image per 45 s; the script queues automatically. A 7-scene video takes about 5–6 minutes.
• Faster: get a key at enter.pollinations.ai and set it as the POLLINATIONS_KEY environment variable — about 1 minute per video, and no watermark.
• Finished images and voice lines are cached, so re-renders are fast.

Limitations (please read)

• This is AI-generated content: visuals are AI stills with camera motion, not true animation. Scripts are AI-written — fact-check anything factual.
• Free-tier images have a small corner watermark; a key removes it.
• Uses stock voices only — no voice cloning, no imitation of real people.
• It does not upload to YouTube, TikTok or Reels; it hands over files and copy.
• Label uploads as "altered or synthetic content" where the platform requires it.
• No guarantee of views or monetisation.
• Relies on free third-party services (Pollinations, Microsoft speech); policy changes could affect availability. Download buyers get free updates for life.

Files

• SKILL.md — agent workflow
• scripts/make_video.py — video pipeline
• scripts/make_thumbnail.py — thumbnail generator
• references/script-guide.md — how to write a short-video script
• references/topic-formulas.md — 10 topic formulas
• references/example_zh.json / example_en.json — example scripts