How do you make a kids counting song cartoon with AI?
By Alverson Layne, Founder
Founder of FXI Studio, building the AI film studio.
Published October 5, 2026 · Updated October 5, 2026
The short answer
Write counting lyrics with one action per line, generate the sung track, and design one mascot as a model sheet. Draw the mascot doing each line in the same set, then give a video model each still plus its slice of the song, so the mascot lip-syncs that line. Cut the clips to the song and add counting captions.
TL;DR
The trick is the slice: each video clip hears only its own line of the song, so the mascot sings exactly those words. Our 32-second demo below took 6 singing shots and about 245 credits for one pass.
Video tutorial
Daisy's counting song: a 32-second AI cartoon made in FXI Studio
The finished demo. Every singing shot was driven by its own slice of the song, then cut back onto the full track so the lip-sync lines up.
Read the transcript
Hello, hello, come and play with me! You're as happy as can be.
Number one, jump up high, jump, jump, touch the sky.
Number two, clap, clap, clap, clap your hands just like that.
Number three, spin around, spin, spin around and round.
Number four, touch your toes, touch your toes and touch your nose.
Number five, wave bye-bye, bye-bye, bye-bye, bye-bye.
Write the song as one action per line
Toddler songs work because they are simple and physical: a number, an action, a rhyme. Give every line one thing the mascot can do on screen, and keep the whole song between 30 and 40 seconds so it fits a short and holds a small child's attention. Here are the lyrics we wrote for the demo. Use them, or swap in your own actions:
"Hello, hello, come and play with me! / Number one, jump up high, jump, jump, touch the sky! / Number two, clap, clap, clap, clap your hands just like that! / Number three, spin around, spin, spin, round and round! / Number four, touch your toes, touch your toes and touch your nose! / Number five, wave bye-bye, bye-bye, bye-bye, bye-bye!"
Every line becomes one shot later, so the lyric sheet is also your shot list.
Generate the song with your exact words
In Create, pick the music model MiniMax Music 2.6 (18 credits a track), leave Instrumental off, and put the style and the lyrics in the same prompt. The prompt we used: "Cheerful children's counting song for toddlers, bright solo ukulele strumming, light hand claps and a soft shaker, one warm, friendly, clear female lead vocal singing slowly and clearly, simple sing-along melody, 112 BPM, major key. Sing these exact lyrics:" followed by the lyric sheet.
Listen before you move on. Our take sang every line in order and added one bonus line of its own ("You're as happy as can be"), which we kept because it scanned. The track ran long past the lyrics, so we used a 32-second stretch from the ukulele lead-in to the last bye-bye. A Lyria 3 Pro take of the same prompt skipped and changed lines: the Lyria models take a prompt only, so they suit instrumental beds better than word-for-word vocals.
SoundStage, FXI's album studio, is built for full records: its songwriter drafts the lyrics with you rather than singing a sheet word for word, so for a nursery song where every word matters, Create is the faster route.
Design one mascot and keep it the same
Design the character once as a model sheet: front, three-quarter, side and back on a plain background. Give it two or three details a viewer remembers and a model can repeat. Daisy is a round baby lamb with cloud-white wool, a peach face, one daisy behind her left ear and a mint-green knitted scarf.
Prompt for the sheet: "Character model sheet for an original children's cartoon, colorful glossy 3D animated style, soft studio light, plain pastel background. A small round baby lamb standing upright on two legs, toddler proportions, fluffy cloud-white curly wool, soft peach face with rosy cheeks, big dark-brown eyes, floppy peach ears, one white-and-yellow daisy behind her left ear, a mint-green knitted scarf. Front view waving, three-quarter view, side profile and back view in one row. Identical design in every pose. No text."
Then save the mascot as a character in your Cast in the Lab, with the sheet as its reference and the same description as its Identity. A tagged character's description is placed word for word at the start of every prompt it appears in, so the wording never drifts between shots. The likeness check is built for human faces; on a cartoon it may say it could not check a take, which does not block anything, so check the mascot by eye.
Make your own character. A well-known cartoon, toy or show mascot is someone else's property, and an original one is yours to build a channel around.
Draw a still for every line, in one set
Write the set once and paste it into every still prompt word for word: "a cozy toddler playroom with a sky-blue wall, a big round window with warm morning sun, a soft rainbow-striped round rug, a low white bookshelf, colorful wooden blocks in a corner and a potted plant." Then change only the action: waving hello, mid-jump with arms up, clapping, twirling with the scarf swinging, bending to touch her toes, waving bye-bye.
Attach the model sheet to each still on Nano Banana 2. Draw one wide still of the whole room too: it opens the film. That is 7 stills in all, and stills are where you fix the pose and framing cheaply, before any video credits.
Make the mascot sing: one song slice per shot
Cut the song into one short audio piece per sung line, in any audio editor, and export each as an MP3. Start each piece just before the first word and end it just after the last. Keep them under 15 seconds: that is the limit for an audio reference.
In Create, pick Seedance 2 Mini. Attach the scene still and the model sheet as reference images, attach that line's audio piece in the audio slot, and set the length to the piece rounded up to whole seconds. Our prompt for the clap verse: "Colorful glossy 3D animated children's cartoon. The little lamb Daisy from image 1 (design exactly as image 2) sings and claps her hands four times right on the beat, in the sunny playroom from image 1. She lip-syncs the sung vocal from audio 1 exactly, mouth shapes matching every word. Static camera, medium shot, gentle bouncy cartoon motion. No text on screen."
Change only the action for each line. Because each clip hears only its own words, the mouth shapes follow them, and the dance moves land on the beat. All six of our singing shots came back usable on the first take. Seedance 2.0 Fast gives a closer, crisper face for three times the credits; use it for the one shot people will pause on.
One rule to know: a clip that uses an audio reference cannot also take a first frame. The still goes in as a reference instead, so the singing shots are separate takes cut on the beat, which is how music videos cut anyway.
Chain the opening move
Open on a camera move into the room rather than a cut. Use P-Video with the wide still as the first frame and the "hello" still as the last frame: "Smooth, steady push-in across the sunny playroom toward the little lamb standing on the rainbow rug; she lifts her hand to wave hello as the camera arrives on a medium shot, ending exactly on the attached last frame. One unbroken camera move, no cut."
Play it at double speed under the ukulele lead-in and it lands right where the first sung line starts. That is frame chaining: the shot ends on the frame the next one begins with, so the film opens as one continuous move.
Edit to the song, then add captions and numbers
Lay the full song on the timeline first and place each singing clip at the moment its slice starts, trimmed to the slice length. Use the full song as the soundtrack rather than each clip's own audio, so the music never jumps between shots and the lip-sync stays lined up.
The editor in the Lab can transcribe the song into captions with word timing and highlight each word as it is sung, with your choice of font, size, colors and position, and the captions are burned into the export. The editor has no dedicated number-badge tool yet: the simple in-studio option is to start each verse's caption with the number ("2! Clap, clap..."), or add number graphics in a finishing pass in any video editor. In the demo we added a big colored circle with the number in the corner in a finishing pass, one color per verse.
Toddlers read color and shape before words, so keep the caption box low and calm, make the current word bright, and let the number be the biggest thing on screen.
What it needs: credits and plan
Singing shots run on Seedance, which needs any credit pack or plan. The free tier makes video with P-Video only, which is enough to try the opening camera move, and the song plus every still comes to 50 credits. New accounts start with 65 free credits and no credit card, about 16 images or 3 free videos (2 today, 1 more next week), plus 22 more credits each week until the free balance is back at 65.
A full pass of this guide costs about 245 credits, inside the 600 monthly credits of The Creator plan ($15/mo).
Publishing for kids on YouTube
If you post the cartoon on YouTube for young children, mark it as made for kids when you upload. That is YouTube's rule for content aimed at children, and it exists to meet COPPA, the US children's privacy law; it turns off comments and personalized ads on that video. Read YouTube's made-for-kids guidance before you publish, and keep the content gentle, original and age-appropriate.
| Step | What you make | Model | Credits (one pass) |
|---|---|---|---|
| 1. The song | One sung track from your own lyrics | MiniMax Music 2.6 | 18 |
| 2. The mascot and set | A model sheet plus 7 scene stills at 4 credits each | Nano Banana 2 | 32 |
| 3. Slice the song | One short audio piece per sung line | None (any audio editor) | 0 |
| 4. Singing shots | 6 clips, 34 seconds in all, at 5 credits a second | Seedance 2 Mini | 170 |
| 5. Opening camera move | One chained shot from the wide still into the first scene | P-Video | 25 |
| 6. Captions and numbers | Word-highlight captions and a counting number per verse | Editor | 0 |
| Total | About 32 seconds of finished cartoon | 245 |
Estimate, not a quote. Prices are the published per-unit credit costs; stills are priced at the 1K size (2K costs a little more). Budget a retake or two on the singing shots. Seedance 2.0 Fast gives sharper faces at 15 credits a second, three times the Mini rate.
How we tested this
We made the demo on this page end to end before writing it: an original mascot (Daisy, a lamb in a mint scarf), original lyrics, one MiniMax Music 2.6 song, 8 Nano Banana 2 stills, 6 Seedance 2 Mini singing shots driven by song slices, one P-Video chained camera move and a caption pass. We also tried a Lyria 3 Pro take of the same lyrics (it skipped and changed lines) and one Seedance 2.0 Fast take of the clap verse for comparison. Including those extras, 2K stills and the tests, the whole run used about 352 credits.
Step by step
How to make a kids counting song cartoon with AI
- 1Write the lyrics
One number and one action per line, about 30 to 40 seconds in all. The lyric sheet is your shot list. Cost: 0 credits.
- 2Generate the song
In Create, use MiniMax Music 2.6 with the style and your exact lyrics in the prompt, then pick the take that sings your words. Cost: 18 credits a track.
Open Create - 3Design the mascot
Make a model sheet of one original character and save it to your Cast so its description goes into every prompt. Cost: 4 credits.
- 4Draw a still per line
Same set, same mascot, a new action for each line, with the model sheet attached. Cost: 4 credits a still, 28 for 7 stills.
- 5Slice the song
Cut one audio piece per sung line, each under 15 seconds. Cost: 0 credits.
- 6Make it sing
On Seedance 2 Mini, attach the still, the model sheet and the line's audio piece, and prompt the action plus lip-sync. Cost: 5 credits a second, 170 for 34 seconds.
- 7Chain the opening
Use P-Video from the wide still to the first scene still for a continuous move into the room. Cost: 25 credits.
- 8Edit and caption
Cut the clips onto the full song, add word-highlight captions and a big number per verse, and export. Cost: 0 credits.
Frequently asked questions
How do you make a kids counting song cartoon with AI?
Write counting lyrics with one action per line, generate the sung track, design one mascot as a model sheet, and draw a still of it for each line in the same set. Give each still and its slice of the song to a video model so the mascot lip-syncs the line, then cut the clips to the song and add counting captions. Our 32-second demo cost about 245 credits.
How do I keep a cartoon character consistent in every AI shot?
Design it once as a model sheet with two or three memorable details, attach that sheet to every still and video shot, and describe it with identical words each time. In FXI Studio you can save it to your Cast so its description is added to every prompt automatically.
Can AI make a cartoon character sing?
Yes. Give the video model the song as an audio reference along with a still of the character, and prompt it to lip-sync the vocal. Slice the song so each clip hears only one line: the mouth shapes follow those words. In FXI Studio this works on Seedance 2 Mini and Seedance 2.0 Fast, with audio pieces up to 15 seconds.
Do I need to know animation to make a kids cartoon with AI?
No. You write the lyrics and describe the character and the room in plain words. The image model draws the mascot, and the video model animates it and matches its mouth to the song. The skills that matter are a simple song, one consistent character and cutting the clips to the beat.
Can AI write and sing a song with my own lyrics?
Yes. In Create, MiniMax Music 2.6 sings lyrics you type into the prompt alongside a style description, for 18 credits a track. Listen to each take: it may add or repeat a line, so pick the take that sings your words.
How long should a toddler counting song be?
Thirty to forty seconds is enough for five numbers, an opening hello and a goodbye. It fits a short-form video and holds a young child's attention, and it keeps the number of singing shots, and the cost, small.
Can I make this on the free plan?
Partly. New accounts start with 65 free credits and no credit card, about 16 images or 3 free videos (2 today, 1 more next week), plus 22 more credits each week until the free balance is back at 65. Free accounts make video with P-Video only, so you can make the song, the mascot, the stills and the opening camera move, but the singing shots run on Seedance, which needs any credit pack or plan.
Do I have to mark AI cartoons as made for kids on YouTube?
If the video is aimed at young children, yes: YouTube requires you to mark it as made for kids, which follows the US children's privacy law COPPA and limits comments and personalized ads on it. Check YouTube's made-for-kids guidance before you publish.
Make your counting song
Start with 65 free credits. Creator plan $15/mo.
Related guides
Hands-on tutorials
- TutorialHow to keep the same character across every shot
- TutorialHow to make a character sheet and turnaround with AI
- TutorialHow to prompt Nano Banana 2 for consistent characters
Reference documentation
Looking for a specific tool? Browse the documentation, the guides, or all tutorials.