Create those children's videos in modelling clay: sculpted characters, plain background, a cut every two and a half seconds. The character sheet keeps the same hero from first shot to last.
This workflow is the real template: the character sheet is mentioned in every shot, the art direction holds the material, the Video node renders ten seconds of stop motion. Explore each node.
At a glance
What the workflow produces
2 images in 16:9 · 1 video of 10s in 720p, 16:9 · 1 voice-over
Canvas structure
9 nodes wired by 5 connections
Models wired in
Seedream 5 Pro, Grok TTS, Seedance 2.0
Cost of one full run
about 132 credits, roughly €1.32
Illustrated clay rhymes are among the most watched formats in children's video, and they rest on a detail a child spots instantly: it is the SAME bird from beginning to end. A character whose beak or colour shifts between shots breaks the story, even when each image is lovely on its own.
This workflow solves that with a three-angle Reference sheet rather than a repeated description. The character is called with an @mention in every prompt: the model receives its actual photos and renders the same creature, shot after shot. The art direction handles the rest, the clay, the thumbprints, the plain background and the soft light.
Seedream 5 Pro carries the character sheet and the material, Seedance 2.0 chains four shots in a single take with that slightly jerky stop motion movement, and OpenAI TTS HD reads the rhyme. The guide on keeping a character consistent explains how to build a good reference sheet.
How it works
1
Sculpt the character
The Reference sheet holds your hero from three angles: front, profile, three quarters. That is what stops it changing face between images, call it with @ in every prompt.
2
Fix the material
The Art direction node describes the clay, the tool marks, the plain background and the macro lens. It feeds every shot, so the world stays coherent even when the scene changes.
3
Write the rhyme
The Text node carries the words, the Voice-over node reads them. A clear voice at normal pace: the audience is four years old and follows the rhythm of the images.
4
Cut four shots into ten seconds
A cut every two and a half seconds, never longer. Describe the four shots in a single video prompt, with the slightly jerky movement that signs stop motion.
Variations and use cases
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
Claymation nursery rhyme
The original use: four shots, one refrain, one endearing character. The rhyme goes into the Voice-over node and each shot illustrates a line. Keep a plain background: busy sets confuse very young viewers and make the character harder to follow between shots.
Stop motion without a camera or a set
The target look is stop motion: fingerprints in the clay, slightly jerky movement, soft tabletop light. Ask explicitly for twelve frames per second in the video prompt, otherwise Seedance 2.0 renders a smooth motion that gives the synthesis away.
Claymation product ad
The product becomes a clay character, the set stays minimal, the voice-over is two sentences. The format works because it is rare in advertising: it catches the eye where a realistic video blends into everything else. See the ten-second video ad for the structure.
Intro for a kids YouTube channel
Five seconds, the character waving, the logo sculpted in clay. Once the Reference card is filled, the same hero comes back identical in every episode: that is what builds a recognisable channel instead of a pile of videos.
Personalised bedtime story
Photograph the child's own soft toy from three angles, drop the photos into the Reference card, and it becomes the hero of the story. The art direction moves it into clay without losing the original colours.
Common mistakes
A Reference card with three identical angles
The model then invents the other side of the head, and the character changes profile in the second shot. Front, profile, three-quarter: three different angles beat ten identical photos.
A set more detailed than the character
In clay, everything around the hero pulls as much attention as the hero. Plain background, two or three props at most, and the eye stays where it should.
A rhyme too long for the number of shots
Four shots cover ten seconds, which is three short lines. The rest of the text needs a second Video node, not a faster read.
Through the Reference sheet, called with @ in every shot. The model receives all three angles on every generation: it does not reinvent the character, it finds it again.
Do I actually have to sculpt something?
No, the character sheet is generated too: three images are enough, made once and reused everywhere. You can also start from a real figurine photographed from three angles.
Can the video run longer than a rhyme?
Yes, by duplicating the Video node per verse and keeping the same sheet and art direction. That is exactly how these channels produce episodes several minutes long.
Do I need to know how to sculpt?
No, nothing is modelled by hand: the Reference card holds generated images of the character from three angles. If you own a real figurine, photograph it and drop the photos in, the render will follow it.
How do I keep the same character across a series?
By reusing the same Reference card from one project to the next: it lives in your library and is recalled with an @mention in any prompt. The guide keeping a character consistent walks through the method.
Is the voice suitable for young children?
Pick a clear voice and slow the pace slightly in the Voice-over node. The template uses a warm voice at 0.92 speed, which carries a nursery rhyme without turning saccharine.
What does a full rhyme cost?
The complete canvas lands around 1.30 euros of credits for ten edited seconds, voice included. A one-minute rhyme therefore costs under ten euros, itemised on the calculator before you sign up.