How to make an explainer video with AI (2026 method)
To make an explainer video with AI in 2026, you start from a one-line topic, a text or your own voice; the tool writes (or transcribes) the narration, splits it into scenes, illustrates each scene with animated elements, adds the voice-over and captions, then renders an MP4. Expect 10 to 30 minutes of human work for a one-minute video, mostly spent reviewing the script. On ClipOnIt, the explainer editor does all of this in vertical 9:16 from the Starter plan ($29/month, 30 explainer videos per month).
What is an "AI-made" explainer video?
An explainer video answers a single question ("why is the sky blue", "how does a mortgage work", "what does our product do") in 30 seconds to 3 minutes, with a voice-over and animated illustrations. "Made by AI" covers three very different levels:
- Script only: ChatGPT or Claude write the text, you edit the rest by hand (Canva, CapCut, After Effects).
- Talking avatar: a virtual presenter reads your text on camera (Synthesia, HeyGen). Handy for training, a poor fit for social feeds.
- Scene-by-scene generated illustration: the tool picks and animates visuals that follow the narration word by word. This is the format of faceless news and science channels on TikTok and YouTube Shorts.
This article covers the third level, the most complete one.
The 6-step method
- A question, not a topic. "Inflation" makes a flat video; "Why your rent rises faster than your salary" makes one people watch to the end. Write the question in one line.
- Narration first. About 140 to 160 spoken English words per minute. One idea per sentence, at most one number per scene, the answer in the first 3 seconds. Let AI write a first draft, then rewrite it out loud: anything you stumble on goes.
- Split into scenes. One scene = one or two sentences, 4 to 8 seconds. A one-minute video usually has 8 to 12 scenes.
- Visuals synced to words. Each element appears when the voice names it. That is what separates an explainer from a slideshow.
- Voice and captions. Synthetic voice or your own; word-by-word captions, because much of the audience watches muted (see the caption styles that hold attention).
- Review and render. Watch the whole video muted, then with sound. Fix the scenes that drag, then export in 9:16, 1080×1920.
How long does it take?
| Step | By hand (Canva, CapCut) | With an AI explainer editor |
|---|---|---|
| Script | 20 to 40 min | 5 to 15 min (AI draft + rewrite) |
| Finding and animating visuals | 1 to 3 h | Automatic, then 5 to 10 min of tweaks |
| Voice-over and captions | 20 to 40 min | Automatic |
| Render | 5 min | A few minutes (about 4 min per video on ClipOnIt) |
The "by hand" durations are orders of magnitude for a one-minute video, not a measurement. The one step that does not compress is judgment: picking the question and rewriting the text.
A concrete example with ClipOnIt
ClipOnIt's explainer video editor accepts three starting points:
- A one-line topic: AI writes the narration, splits it into scenes and composes each scene.
- A text you already wrote: it is read as is.
- Your voice: you narrate into the mic, in one or several takes, or import an audio file. The narration is transcribed word by word, split into scenes, and every element is synced to your words. You can re-record a single sentence.
Illustrations come from a library of about 11,000 animated elements in two styles: Pixel News (2D pixel art) and Archibald (low-poly pseudo-3D with a moving camera). The synthetic voice offers 11 voices. Then three ways to edit:
- The AI Director: you type or dictate "shorter", "a chart here", "change the color"; each proposed change is explained in one sentence, accepted or rejected, and everything can be undone.
- Variants: three other versions of a scene, side by side, keeping the narration; the cost is shown first.
- By hand: move, hide or lock an element on the canvas and timeline.
The output is a vertical MP4 ready to post. Version history keeps every state.
What an AI explainer video costs on ClipOnIt
| Plan | Price | Explainer videos / month | AI Director requests / day | Voice minutes transcribed / month |
|---|---|---|---|---|
| Free | $0 | First try on an example | — | — |
| Starter | $29/month ($290/year) | 30 | 15 | 60 |
| Pro | $49/month ($490/year) | 100 | 40 | 300 |
| Business | $149/month ($1,490/year) | 400 | 120 | 1,200 |
Source: ClipOnIt plan limits as of October 29, 2026, see the pricing page. The same subscription includes video clipping (20, 60 or 200 long videos per month).
The most common mistakes
- A script that is too long. Past 90 seconds, every sentence must earn its place.
- Decorative visuals. If the image does not show what the voice says, it distracts.
- Publishing the AI's first draft. Generated text is correct but generic; the video gets interesting when you add a verified number, a real example or an opinion.
- Unchecked facts. AI can get a date or a figure wrong. Check every number before rendering.
FAQ
Can ChatGPT make an explainer video?
ChatGPT writes the script, not the video. To get to an MP4 you need an editing tool; ClipOnIt also exposes its explainer editor through an MCP server, so Claude or ChatGPT can produce the video end to end (the explainer_produce tool).
Do I need to draw or animate?
No. Elements are already drawn and animated; AI picks and places them. You can draw a pixel-art element yourself if the catalog lacks one.
Can I use my own voice?
Yes: record with the mic or import an audio file, transcribed word by word. Transcribed minutes are capped per plan (60 to 1,200 per month).
Which format for TikTok, Reels and Shorts?
Vertical 9:16, 1080×1920, MP4; that is the editor's output format. Lengths that perform are covered in short video length and format in 2026.
What about 16:9 for YouTube?
Not today: ClipOnIt's explainer editor renders vertical only. For animated 16:9, see the explainer video tools comparison.
Written by the ClipOnIt team. Updated October 29, 2026.