Make a bedtime story video for kids with AI

Children’s animation is where an AI video pipeline stops fighting its own limits. Stylised characters tolerate drift, stories are short, there is no dialogue, and one warm narrator is the natural form.

Why cartoons work where photoreal does not

A cartoon fox with an oversized red scarf reads as the same fox across shots even when its proportions shift a little. A photorealistic human face does not — the eye is unforgiving about people and generous about drawings.

So the technique is not to fight the drift. It is to anchor on something that survives it.

One shape, one colour, one prop — repeated word for word in every scene.

Why this runs through Claude, not a form

A film is written scene by scene, with questions asked before a word is committed. That is a conversation, not a web form with a dropdown. So the app connects to Claude as an MCP connector — one URL, added once, no key to paste:

Add it under Settings → Connectors → Add custom connector, sign in when prompted, and Claude gains ten tools: fetch the production guide, see what each option costs, create a video, check its progress, list its scenes, drop a scene, set the end card, re-cut, list your uploaded images, and join finished videos into one longer film.

From then on you ask in plain sentences. Claude reads the production method first, asks what it needs, writes the script to the right length, and renders it — then edits it with you afterwards, for free, because re-cutting reuses clips that already exist.

That is the difference from a trailer generator you fill in: the thing that plans the film is the same thing you are talking to, and it can go back and change one scene without rebuilding the rest. More on automating video with Claude →

Start with the character, not the story

Before a word of script is written, Claude generates one portrait of the character and shows it to you. If it is not right, the sentence changes and it tries again — one image, not fifteen. Once approved, that picture anchors every scene the character appears in.

Getting this wrong and finding out after fifteen scenes is the expensive mistake. One approval up front prevents it.

Age changes everything

• 2–4 — one idea, heavy repetition, 45–60 seconds. No jeopardy at all: the “problem” is a puddle in the way. Repetition is the pleasure, not a flaw.
• 5–7 — a problem, two failed attempts, then success. 60–90 seconds.
• 8–11 — more plot and a mild setback allowed. 90 seconds to three minutes.

The shape of a children’s story

Not three acts. This:

Meet → Want → Problem → Try, try again → Put right → Home

Nothing frightening, no darkness left unresolved, and the ending is always safe and warm. Tension in a children’s story is a lost mitten, not danger.

Write slower than you would for adults

About 1.5 words per second, not two. A 60-second story is roughly 90 words. Sentences of eight to twelve words, one idea each, concrete words throughout — “the scarf caught on a branch”, not “an obstacle presented itself”. If it does not sound like a bedtime story read aloud, it needs rewriting.

This one needs AI imagery

There is no cartoon fox in a stock footage library, and asking for one returns a photograph of a real fox. Children’s stories use generated illustration, which spends your own Pollen — roughly one generation per scene. Claude will say so before generating anything.

Watch it before a child does

Generated pictures are unpredictable in a way a script is not. Occasionally one comes out odd or off-tone even when the words are exactly right. An adult viewing it first is the only real safeguard — and if a scene is wrong, dropping it re-cuts the film for nothing.

Next: make a cinematic trailer, or see everything Claude can do from chat.