Vox Media publishes videos online to help explain some of the chaos happening around us. From science to economics, Vox's YouTube videos help viewers understand common phenomenon that impact human behavior and society.

These videos are particularly popular for their collage-style animations. While these effects look like they require hours of work in After Effects. However, AI image and video generators have made it possible to recreate a similar documentary aesthetic using a much simpler workflow. Below is the video that we created with the following tutorial:

In this guide, we'll walk through how to make Vox-style animations with AI video generation, from writing a script to generating animated paper collages and editing everything into a finished video. We'll also explain how to keep the visual style consistent across scenes and adapt the workflow for different types of stories.

Table of Contents:

• What Is Vox-Style Animation?
• How to Make Vox-Style Animations With AI
• Frequently Asked Questions

What Is Vox-Style Animation?

Vox-style animation is a form of documentary motion graphics that uses layered visual elements to explain a story. Instead of relying entirely on traditional footage, creators combine photographs, maps, newspaper clippings, illustrations, and typography into animated compositions.

Common visual characteristics include:

• Layered paper cutouts: Photographs, documents, and graphics that appear physically cut out and arranged on top of one another.
• Archival textures: Newsprint, halftone photography, faded maps, and visible paper grain.
• Limited color palettes: Desaturated beige, black, and gray, with occasional red or yellow accents.
• Stop-motion movement: Elements slide, drop, or snap into position rather than moving with smooth, continuous animation.
• Documentary narration: A voiceover carries the story while each animated scene illustrates a specific detail.

Although Vox’s video producers are some of the best in their fields, amateur video creators can emulate this production style with accessible editing techniques.

How to Make Vox-Style Animations With AI

The workflow involves six main stages: writing the script, dividing it into scenes, generating the images, animating them, creating a voiceover, and assembling the final video.

Step 1: Choose a Topic and Write Your Documentary Script

Start by choosing a documentary topic. Historical events, unsolved mysteries, biographies, and major news stories are good candidates because they naturally involve dates, locations, people, and documents that can become visual elements.

For this workflow, we used a detailed master prompt, which turns an AI chatbot into a documentary production assistant. Paste the master prompt into ChatGPT, Claude, or Gemini.

The chatbot will first ask you to upload a source-material PDF containing additional references or type skip to use the built-in instructions.

You can download the source-material PDF below and upload it.

Next, you'll choose a niche from categories such as crime, history, technology, sports, or mysteries. The chatbot will suggest ten documentary ideas before asking you to select a topic and a video length.

Once you've selected a duration, the AI will write a complete script. Review and edit the generated script before continuing.

Step 2: Break the Script Into Visual Scenes

Once your narration is finished, the next step is to divide it into smaller sections called visual beats.

Each beat represents one idea in the narration that can be illustrated with a single image. Rather than generating a continuous documentary animation, you'll create individual scenes to combine later. A short documentary might contain 12–15 beats, while a five-minute documentary could contain more than 100.

The chatbot will break the narration into beats of approximately two to three seconds, or five to eight spoken words. It also generates a table containing the beat number, starting timecode, and exact narration words.

Review this breakdown before generating your images. Each beat should communicate one visual idea rather than trying to illustrate an entire sentence containing several different events.

Step 3: Generate the Paper-Collage Images

Now you'll generate one image for every visual beat. These images will become the finished compositions that your AI animations build toward.

The master prompt already contains the exact image-generation instructions, so you don't need to write each prompt manually. The chatbot will export the completed prompts into a .txt file, with one prompt per scene and blank lines separating them.

Open Kai and paste these image prompts into the prompt box. You can only generate 4 images at once, so you'll need to do this in batches.

Make sure to label each image by its order, to not get confused during later steps.

Step 4: Animate Each Collage Image With AI

Now that you've generated all your paper-collage images, the next step is to turn each one into a short animation.

Instead, you'll animate each image separately using the prompt below, then combine the finished clips during editing.

Check that the paper moves naturally, the camera stays stationary, and the animation ends on the reference image.

Then, repeat for the remaining images.

Step 5: Generate the Documentary Voiceover

With your scenes ready, create the narration. The master prompt includes a dedicated voiceover stage using ElevenLabs. To generate the narration, return to your chatbot and type: voice

The chatbot will prepare a copy-paste version of the script along with the recommended voice settings.

Copy and paste this prompt into Kai to generate a voiceover.

Make sure to suggest ElevenLabs as the model.

Step 6: Edit the Animations Into a Finished Documentary

The final step is to assemble your animated clips and narration into a complete video.

Open the Studio and upload the generated animation clips and your voiceover, then arrange the scenes in the order specified by your visual beat breakdown.

Use the narration as your timing reference. Each animated collage should appear when its corresponding detail is mentioned.

Because the generated animations are six seconds long, some clips will need to be shortened significantly. You can trim the beginning or end of each animation, depending on whether the paper assembly or the finished composition is more important to that moment.

Once the timing is set, add the finishing touches:

• Background music: Choose a subtle documentary soundtrack that supports the narration without overpowering it.
• Sound effects: Paper shuffles, stamp impacts, typewriter clicks, and other tactile sounds can reinforce the stop-motion effect.
• Subtitles: Use automatic subtitles to make the documentary easier to follow, especially on social media.
• Transitions: Favor simple cuts over elaborate transitions. The paper animations already provide much of the visual movement.

Frequently Asked Questions