Trying Cuddler: An AI Story App for Video and Characters

I recently tested a new app called Cuddler. It's an AI story app where you can explore characters, watch stories, and create your own videos and comics. I wanted to share what it offers and a few ways to get started with AI storytelling.

Cuddler calls the idea Living Stories. You might watch a scene, become curious about someone in it, and continue through a conversation with that character. A moment from that conversation can also become material for a new video. That connection between watching, talking, and making is what I want to explain here.

There are several ways to use Cuddler. I'll walk through the main ones and show where I would start with a small project.

A quick overview: you can explore existing stories, develop a character through conversation, or go straight to Video Studio with an idea. You don't need to use every feature to make something.

Here is Cuddler's official promotional video, posted by @CuddlerAI.

// Detect dark theme var iframe = document.getElementById('tweet-2099786672904999417-628'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=2099786672904999417&theme=dark" }

Start an AI story with a character

The Explore page introduces characters with different personalities and opening situations. You can choose one or create your own, then write what you say or do as the conversation develops.

For me, the useful way to explain this is through a small example. Suppose I want a story about a detective meeting someone who knows more than they admit. I would start with that situation and a few lines of dialogue. Before worrying about camera movements, I want to know what the detective wants and what the other person is hiding.

That's an example of a starting point, not a finished script. Conversation gives you a way to develop it. Cuddler also has tools for giving characters a voice and turning story moments into images or video. The character and storytelling overview explains how those parts connect.

If you would rather see what people have made first, the Stories page is another place to start. You can browse published work without beginning with a blank prompt.

Turn an AI story idea into a video

You can also skip character chat and head straight to Video Studio.

Studio has a chat mode for working on individual images and clips. You describe what you want, add references when needed, choose from the available options, and review the price before generating. You can keep an image as a reference for a later video, which is useful when a character's appearance matters.

For a first attempt, I would keep the idea small. Here is a prompt I would use as a starting brief:

A detective waits in an empty cafe after closing. Someone places an envelope on the table. The detective recognizes the handwriting and stops reaching for it. Keep the camera close enough to see the reaction.

That gives the shot a place, an action, and a change. I would look at the result before adding more people, movement, or dialogue. Does the envelope arrive clearly? Can I understand the reaction? Those questions are more useful than making the prompt longer just for the sake of it.

Build a film with several shots

For a longer idea, Studio also has Director mode. It takes a brief through a shot-by-shot screenplay, reference art, rendering, and an assembled cut. Individual shots can be regenerated when they need another attempt. The Studio walkthrough covers that process.

I would make the first brief specific about the cast, setting, and ending. For the detective example, that might mean three shots: the empty cafe, the envelope arriving, and the detective deciding whether to open it.

Before rendering, I would check whether each shot adds something. Two shots that both show someone waiting may only need to be one. I would also settle the character's clothes and the room's appearance in the references, so there is something concrete to match across the sequence.

Cuddler's Studio introduction also describes reusable workflows called skills, including character introduction reels and turning a chat scene into a video. These give you starting structures for common tasks.

A few creator videos to look at

Alongside the official introduction, here are three creator posts. I've linked the original posts so you can see each creator's own description and ask them about their process.

Posted by @RotimionX

// Detect dark theme var iframe = document.getElementById('tweet-2112077995552747648-83'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=2112077995552747648&theme=dark" }

Posted by @ottleyai

// Detect dark theme var iframe = document.getElementById('tweet-2108960755076956438-631'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=2108960755076956438&theme=dark" }

Posted by @Creator_arti11

// Detect dark theme var iframe = document.getElementById('tweet-2108142734506913810-700'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=2108142734506913810&theme=dark" }

Let the audience make a choice, or tell it as a comic

A film is one way to tell the story. Cuddler also supports interactive AI stories with branching choices, as well as comics.

For the detective scene, I might want the audience to decide whether to open the envelope or follow the person who delivered it. That is where an interactive structure becomes useful: the choice should lead somewhere different.

A comic would call for a different treatment. I could spend a panel on the handwriting, another on the detective's face, and let the reader decide how long to pause. I like having those options around the same story idea, without feeling that every idea needs to become a video.

A note for developers

Since I'm sharing this on DEV, I also want to mention the Cuddler API. It provides access to video, image, speech, and music generation for people building their own applications.

Generation runs as asynchronous jobs: submit a request, then retrieve the result or receive a webhook. That's a separate way to use Cuddler from making something in Studio. The API introduction links to the quickstart and explains the job flow.

What I would keep in mind

I wouldn't approach AI video expecting every first result to be usable. Reference images help guide a character's appearance, but I would still check faces, clothing, movement, and continuity between shots. A small test is a sensible place to start before committing to a longer project.

I would also check the current model options and cost inside the product. Availability and limits can change, and the Studio and API have different billing arrangements.

The part I find most interesting about AI story creation in Cuddler is being able to stay with an idea as it changes form. A conversation can help develop a scene. A scene can become a clip. A different choice can suggest another direction for the story.

If you're curious, you can browse Cuddler or start a small project in Studio. I'd be interested to hear what you'd try first, and which part of the process you would want me to explain in more detail.