For creatorsMake videos about AI coding? Earn 50% for life.Earn 50% for life.Become a Morphy creatorJoin →
morphy
Download

POV horror 3D

Mannequin

A free Claude skill for POV horror stories

It tells a story to "you", slowly. You miss the last train, the gates slam shut, your phone is dead, and a note in the driver's booth has one rule. You shrug it off. Then, exactly at midnight, every light goes out, a silent train pulls in and something starts breathing on the back of your neck. You have two choices. It all happens in a wet, hazy London-style subway station at night, acted out by a glossy white artist's mannequin with a painted face and a black one with two glowing eyes. Hard cuts every couple of seconds, each a slow push-in, white caption chunks that go quiet in the pauses, and an ending where 1 GET ON and 2 STAY draw their lines to the open door and to the dark, so the comments fill with answers.

30–100 s9:16 verticalTikTokInstagram ReelsYouTube Shorts

Download mannequin.mdReference frames

Free · by Casper

What it makes

  • A glossy white artist's mannequin with a painted face: annoyed, sad, smug, scared, terrified
  • A wet, hazy subway at night: white tiles, navy signs, red seats, a NEXT TRAIN board, a wall clock, a silent train
  • Story beats on their word: a gate that slams with a white flash, a clock that flips to 00:00, lights that die far to near
  • A black mannequin with glowing eyes that comes up behind him, a step at a time
  • Two choice pills, 1 GET ON and 2 STAY, whose lines draw to what they mean, then a long push into his eyes

Best for: Choose-your-path stories that get people answering in the comments: a creepy what-would-you-do, a dilemma, a rule you shouldn't break.

Twelve frames from a finished Mannequin reel, in order
Twelve frames from a finished Mannequin reel. The agent gets the same picture to match.

What you need

  • Node 22 or newer
  • Chrome or Edge
  • ffmpeg
  • A graphics card
  • three.js (the agent installs it)

30 to 100 seconds · Narrated by ElevenLabs, a free neural voice or your own recording, timed word by word · Its own synthesized subway sounds, about 20 minutes to draw a 99-second reel

How to use it

  1. Morphy: it's built in. Open Settings → Community → Playground → Skills, kind Content, and add Mannequin to a chat in one click.
  2. Claude Code: save the file as ~/.claude/skills/mannequin/SKILL.md, with the reference frames beside it as reference.jpg, then ask for a video or run /mannequin.
  3. Codex or any other agent: add it as a skill, or attach the file to your first message.

Your first video

Make a 60-second Mannequin reel: you fall asleep on the night bus and wake up at a stop that isn't on the map. End on two choices.

The agent checks your setup, writes the script, makes the voice, renders a few stills to check and then every frame, and hands you the MP4. Posting it is up to you.

Questions

Is the Mannequin format really free?
Yes. Download the .md file: no account, no email, no paywall. Use it for as many videos as you like.
What do the voices cost?
The voice comes from ElevenLabs. Its free plan is enough to try the format (about 10,000 credits a month, roughly one per character of script). Posting videos that promote anything needs a paid plan for the commercial licence, from $6 a month.
Does it work with Codex too?
Yes. It's a standard SKILL.md, the skill format Claude Code and Codex both read, and Morphy has it built in. Any other agent that can run code can follow it if you attach the file.
Do I have to label the video as AI?
On Instagram, a reel narrated by a realistic AI voice needs Meta's AI info label; switch it on when you post. On YouTube and TikTok, check the AI disclosure setting when the content looks realistic.

More free formats

All formats in Library.md
Download Morphy