October 30, 2026
How to make an explainer video with AI, step by step
Make an explainer video with AI. Write the script, make a picture for each scene, add a voice and captions, and export a finished video in genvi.
An explainer video is a short film that explains one idea. You can now make one with AI, without a camera or a crew. This guide shows how to make an explainer video with AI in genvi, from the script to the finished file.
What you will make
A short video of 45 to 90 seconds that explains one thing, such as how your product works. It has pictures or clips for each scene, a voice that reads the script, and captions on the screen.
What you need
- One idea to explain, in a sentence. If you cannot say it in a sentence, the video will be confusing.
- A genvi account with credits.
- About an hour for a first video.
Want the fastest route? The Shorts format turns a topic into an editable short video in one go. See how to turn a topic into a short video. Use the steps below when you want full control of each scene.
Step 1: Write the script
The script is the words the voice will say. Write it first, because everything else follows from it.
- Use short sentences. Say one idea at a time.
- Start with the problem the viewer has.
- Then show the answer in three steps or fewer.
- End with one thing for the viewer to do.
People speak about 2.5 words each second. So 60 seconds is about 150 words. If you want help with a draft, open Chat Studio and set Kind, the type of result, to Text. Ask for a script for your idea, and ask it to keep sentences short.
Step 2: Split it into scenes
A scene is one picture or clip on screen. Change the scene every one or two sentences, so the video keeps moving. For each scene, write a line about what we should see.
A small table makes this easy.
| Scene | What the voice says | What we see |
|---|---|---|
| 1 | Most plants die from too much water. | A sad plant in a puddle |
| 2 | Check the soil with a finger first. | A hand testing soil |
| 3 | Water only when it feels dry. | A watering can, a happy plant |
Step 3: Make a picture for each scene
You can make the scene pictures in two places.
- Chat Studio. Describe the picture, pick an image model and send. Good for a few scenes. You can read more in Chat Studio: describe it, get it back.
- Node Studio. Add one image block per scene on the canvas and run them together. Good for a longer video.
The big challenge is keeping the same look in every scene. Three habits help.
- Pick one style and copy it into every prompt. For example, "flat colourful illustration, soft light, plain background".
- Use one model for all scenes. Different models have different looks. To pick one, read how to choose an AI image model.
- Save a character or mascot as a reference. In genvi you can save a reference once. Then type its handle, which starts with @, in any prompt. The same description and pictures are used every time.
A paper-craft scene made with FLUX 2 Flash. A single material like paper gives every scene in a video the same feel.
A city scene made with Grok Imagine 2.0. A wide opening shot like this is a good start for an explainer about a big idea.
Step 4: Bring some scenes to life
A video made only of still pictures can feel flat. You can turn a picture into a short clip with a video model. Do this for the important scenes only, since clips cost more than pictures. Read how to animate a photo with AI for the steps, and how to write prompts for AI video for what to say.
For the other scenes, you can simply keep the picture on screen for a few seconds.
An illustrated scene made with FLUX 2 Flash. Warm, simple pictures like this suit an explainer that tells a small story.
Step 5: Make the voiceover
You can make the voice in Chat Studio or Node Studio. In Chat Studio, set Kind to Voice, paste your script and pick a voice model. In Node Studio, connect a Text block to a Text to Speech block.
genvi has many voice models, such as ElevenLabs v3, OpenAI TTS-1, Gemini voices and Kokoro. Kokoro is the cheapest, so use it to check the timing. When the script is final, make the last version with the voice you like most. Listen to the whole thing. If a word sounds wrong, spell it the way it sounds and run again. The steps are in how to add AI voiceover and captions.
Step 6: Put it together in the Video Composer
The Video Composer is genvi's video editor. Your pictures, clips and sounds sit in rows called tracks, and the tracks play together.
- On the Dashboard, click New Project and choose Video Composer.
- Drag the voice file onto an audio track.
- Drag your scene pictures and clips onto a video track, in order.
- Drag the edge of each one so it lasts as long as its line of the script.
- Play the whole thing. Move scenes until the picture matches what the voice says.
Step 7: Add captions
Right-click the voice file and choose Auto Captions. genvi listens to the voice and writes the words on a new Captions track. Each caption shows up to six words. Captions are normal text, so you can fix a word, change the colour or move them. Many people watch with the sound off, so captions matter.
Step 8: Export
In the top toolbar, click the small arrow next to Save and choose Export video. Choose a resolution (720p, 1080p or 4K) and a shape: 16:9 for YouTube and your site, or 9:16 for phones. Then click Start Export. The export runs in your browser, so keep the tab open until it finishes.
Tips for a clear explainer
- One idea only. Cut anything that does not help it.
- Show, then tell. Put the picture on screen a moment before the words.
- Keep scenes short. Three to six seconds is plenty.
- Add a little music, quietly. Turn it down so the voice stays clear.
- Watch with the sound off. If the video still makes sense, the captions are doing their job.
What it costs
An explainer is a mix of runs: pictures, a few clips and a voice. Each has its own price, and genvi shows the price before you run it. Failed runs give your credits back. Auto Captions and the export do not use credits. To keep costs down, draft with cheaper models first, and animate only the scenes that need to move.
FAQ
How long should an explainer video be?
Most work well between 45 and 90 seconds. Shorter is better if the idea is simple.
Can I use my own voice?
Yes. You can upload your own recording to your files and drag it onto an audio track. Auto Captions works on any clip that has speech.
How do I keep the characters the same in every scene?
Use one model, one style line in every prompt, and a saved reference for your character. For video clips, read how to keep a character consistent in AI video.
Do I need to edit video before?
No. If you can drag a picture onto a row, you can use the Video Composer.