Guides

How to Make a 3D Explainer Video with AI, from Your Documents and a Chat

To make a 3D explainer video with AI, decide your goal, audience, and length, attach your documents, and settle the storyboard in chat before moving on to the build, preview, and final export. This guide covers what to prepare, the five steps, a storyboard checklist, sample requests, and how to fix common problems.

Published 2026-11-04 · Updated 2026-11-07 · 3DCG Video AI

Key points

  • Making a 3D explainer video with AI takes 5 stages: discuss → check the storyboard → build and check stills → preview → final export.
  • Before you start, decide on the purpose, audience, length, and aspect ratio, and gather materials that show the numbers and facts. Your storyboard will come together faster.
  • The most important step is checking the storyboard. Before you move on to build, review the rough sketch, structure, and narration script for each scene.
  • In Standard mode, for a 1-minute video, the storyboard takes a few minutes, the build takes about 30 minutes to 1 hours, and the final export takes about 1 hours. Processing continues even if you close the page.
  • As a guide, a 1-minute video with 2 small revisions costs about 3,500 credits (about $35).

What to decide before you start

Purpose and audience

Decide in one sentence what you want viewers to do or understand afterward, such as "explain what the product is at a trade show" or "help new staff understand why the inspection steps matter." The wording and level of detail change depending on whether viewers are experts or learning about it for the first time.

Length

Choose a length from 15 sec to 10 min. The video's length is set by the narration, so the final result may differ from what you specify (typically by about ±10%). For clips that loop on social media or at trade shows, 15 sec to 1 min works well. For product or how-it-works explanations, try 1–3 min. For training, 3–5 min is easy to work with.

Aspect ratio

Use 16:9 for websites, presentations, and widescreen monitors, 9:16 for short videos on smartphones and vertical displays, and 1:1 for social media posts. Each video is exported in one aspect ratio, so decide where you'll use it first.

Materials

The AI builds the structure using only the numbers and facts in your materials. If you gather materials that cover what you want to say, the storyboard will be more specific. You can attach these file types:

  • PDF, Word, PowerPoint, Excel (OpenDocument also supported)
  • Images (PNG, JPEG, WebP, GIF)
  • Text (TXT, Markdown, CSV, JSON)
  • CAD 3D models (STEP, IGES, STL, OBJ, PLY, glTF/GLB, 3MF) and drawings (DXF, DWG). 3D models can be used as is in the video

Files can be up to 30 MB each, and the documents the AI reads can total up to 20 MB per video (3D models are counted separately). Export proprietary formats such as SolidWorks to STEP, and Jw_cad drawings to DXF, before attaching them. DWG files can be read as they are, but tables and some shapes may not be read, and hidden layers also show up in the drawing, so we recommend exporting to DXF and attaching that. 3D solids in DWG files don't appear in the drawing, so attach a STEP file when you want to show a solid shape.

5 steps to make a 3D explainer video with AI

Step 1: Chat and attach your documents

After you sign in, tell the AI in chat what video you want to make and attach your documents. If you write the purpose, audience, length, aspect ratio, and voice style all together, the first storyboard will be more specific. You can choose the length below the chat input box. You sign in with an account such as Google, and we don't store your password.

Step 2: Check and revise the storyboard

The AI (Anthropic's Claude) reads your materials and, within a few minutes, proposes a storyboard that pulls together a rough sketch, structure, narration script, and presentation for each scene. The rough sketches are for checking composition. The finished video is built in 3D with this layout. Tell the AI in chat what you want changed, and it revises the storyboard. Settling the content at this stage, before the build, is the surest way to cut down on rework.

Step 3: Build and review the still images

When you move on to Build, the AI creates 3D models, graphs, diagrams, and camera work in 3D. Before each run, we show the estimated credits and the maximum. If you want to use High Quality mode, choose it before you build. A 1-minute video in Standard mode takes about 30 minutes to 1 hours, and you can check the result in still images. If anything about the shapes, colors, or text looks off, tell us in the chat.

Step 4: Check the flow and voice in the preview video

In Export, we first create a lightweight preview video. Watch the AI narration (synthetic voice from Google Gemini), captions, and scene transitions all the way through, and check the speaking speed and how the scenes change. It also helps you spot mistakes in how numbers and proper names are read aloud, and errors in the captions.

Step 5: Export your final video and keep it

If the preview looks good, export the final 1080p MP4. A 1-minute video in Standard mode takes about 1 hours. Processing continues even if you close the screen. Finished videos are deleted automatically after 30 days, so choose "Keep" if you want to save one. You can keep a video for 6 months, 1 years, or 3 years, paid up front based on file size. For a 1-minute video, keeping it for 1 years costs a few to about 10 credits.

Storyboard review checklist

  • Does the first scene make clear what the video is about?
  • Is the story in the order viewers want to learn it?
  • Do the numbers, units, model numbers, and proper nouns match your materials?
  • Is there anything that isn't in your materials, or any overstated claims?
  • Do difficult terms have a short explanation for viewers?
  • Whether the rough sketch for each scene matches your vision for the main subject's position and size and the text placement
  • Whether how each scene presents things (what appears in 3D and where the camera moves in close) fits what you want to convey
  • Whether the amount of narration fits the length you specified
  • Whether the aspect ratio and voice tone suit where you'll use the video
  • Does the last scene convey a summary of the key points and what viewers should do next?

Example requests

When you make a request, the trick is to include who it's for, what it's about, how long it should be, and the aspect ratio and mood you want, all at once.

Example 1: Explain how a product works (16:9, 90 sec)

"Based on the attached catalog and spec sheet, make a video that explains how our new heat exchanger works in 90 seconds, for engineers visiting our trade show booth. Use 16:9, and show in 3D how 2 fluids pass heat to each other inside. Use the performance numbers exactly as in the documents, and use a calm voice."

Example 2: Lesson introduction (16:9, 1 min)

"Based on the attached lecture materials, make a 1-minute video to open a chemistry lecture for college students in year 1. I want to show in 3D, while rotating, how sodium and chlorine are arranged inside a salt crystal. Please add a short explanation for technical terms."

Example 3: Short video for social media (9:16, 30 sec)

"Narrow the new product's features down to 3 and make a 30-second vertical video (9:16) for social media. Make it clear what the product is in the first 3 seconds, and use large text. Use a bright, energetic voice."

You can also find examples of how to make requests on the Product introductions, Trade shows and sales, Training and manuals, and CAD data pages.

How to fix things when they don't go well

  • The content doesn't match your materials: Say which part of which scene should match which part of your materials (example: "Match the values in scene 3 to table 2 in the spec sheet"). If your materials are missing something, attach more.
  • The length doesn't match your request: Say "Make it 60 seconds," and the AI adjusts the amount of narration and revises the storyboard.
  • The presentation doesn't match what you pictured: For each scene, say what you want to show and how the camera should move, such as "Move in close to the part in this scene" or "Circle around the whole thing."
  • The shape differs from the real thing: Add photos or drawings, or attach a CAD 3D model to use that shape as is.
  • Colors or text are hard to read: Look at the still images and say things like "Make the colors in scene 2 brighter" or "Make the text bigger."
  • The voice doesn't fit: Say something like "a calm voice" or "a bright, energetic voice" to choose a different voice style.
  • You want a major rework: Tell the AI in chat what you want changed, and it can redo the work starting from the storyboard.

Each revision uses credits for processing. Send your notes together in one message to keep the number of revisions and credits down.

FAQ

Do I need to know 3D software or video editing?

No. The AI operates the 3D software. You tell it what you want in the chat, review the storyboard, still images for review, and preview video, and say what you'd like changed.

How long does it take to finish a video?

In Standard mode, for a 1-minute video, the storyboard takes a few minutes, the AI's 3D build takes about 30 minutes to 1 hours, and the final export takes about 1 hours. High Quality mode takes several hours each for the build and the export. Processing continues even if you close the page.

How is this different from using video generation AI?

Unlike video generation AI that creates video directly from pixels, 3DCG Video AI assembles objects in 3D before rendering them, so shapes, text, and numbers are much less likely to break down. It uses the shape of your CAD 3D model as is, and when you want to make changes, you can rebuild from the storyboard. The rough sketches in the storyboard are for checking composition; shapes and dimensions come from your CAD data.

How much does it cost?

A 1-minute video built with 2 small revisions is about 3,500 credits (about $35). Each big revision, such as changing the background, adds about 0.6–880 credits. High Quality mode costs roughly 0.6–7,100 credits for 2,500 minutes. Before you build or export, you'll see an estimate and the maximum. For a comparison with production company rates, see the cost of 3D explainer videos and pricing when you use AI.

Can I use my videos commercially?

Commercial use is allowed within the Terms of Service, on condition of payment. The narration is an AI-generated voice.

Are attached materials used to train the AI?

No. Attached documents are used only to build your video and, under each AI provider's business terms, are not used to train AI. Documents and results are deleted automatically after 30 days.

Check the result with sample videos

There are 51 sample videos made with the same AI and 3D software as this service (engineering, chemistry, physics, business, product introductions, and 1 in High Quality). Check how 3D shapes, narration, and captions look and feel in How a 4-stroke engine works and the High Quality jet engine (turbofan), among others. Sign in and you can get started right away.

This article was researched and written with AI, using published information and this service's own measurements. If you find a mistake, email service@3dcgdouga.com. Who runs this site is on the Legal Notice page.

Turn your documents and chat into 3D animation

The AI plans a storyboard, builds 3D visuals, and turns them into a video with narration and captions. You pay only for what you use.