AI motion graphics generator: how to make motion graphics with AI, step by step

To make motion graphics with an AI motion graphics generator, lock your script first, turn it into a timed scene list, prompt one scene at a time with exact text, layout, motion, timing and brand details, then edit instead of regenerating. Add voice and music, export in the right formats and run a short quality check. The generator speeds up the animation; the planning and checking around it decide whether the result is any good.

This guide walks through the whole workflow we'd use for a 30 to 90 second piece: a product explainer, a feature announcement, a data story. It works with any of the common tools. If you haven't picked one yet, our comparison of AI motion graphics tools by job helps you choose.

Rather skip the eight steps? We make animated explainers for businesses every day. Paste your website link and get a watermarked draft within 24 hours. Three revision rounds are included, and to keep it you subscribe from $99 a month plus sales tax.

Before you start: the five decisions that save the most time

Settle these in ten minutes before you open any generator. Every one you skip comes back later as rework.

1. One message. Finish this sentence: "After watching, the viewer understands that ___." If you need two sentences, you need two videos.
2. One action. What should the viewer do next: book a demo, sign up, buy, read the docs?
3. Length. 15 to 30 seconds for social, 60 to 90 seconds for a homepage or product explainer. Our guide on how long an explainer video should be goes deeper.
4. Format. 16:9 for websites and YouTube, 9:16 for Reels, Shorts and TikTok, 1:1 or 4:5 for feeds. Decide now; reframing later is real work.
5. Brand kit. Two to four colors as hex codes, one or two fonts, your logo as SVG or a transparent PNG. Generators that receive these produce graphics that look like you, not like a template.

Step 1: Write the script and lock it

Motion graphics follow the words, so the words come first. Even a video without voice-over needs a script: it's the on-screen text, in order.

• Write for the ear. Short sentences, one idea each. Read it aloud and cut anything you stumble over.
• Count the time. Voice-over runs at roughly 140 to 160 words per minute, so a 60-second video holds about 150 words. Most first drafts are twice that long.
• Use AI as a sparring partner, not an author. A language model writes a fast first draft. It also writes generic sentences that could describe any company. Replace every generic claim with something specific to you: a number, a customer's situation, a real feature.
• Lock it. Get sign-off from whoever has to approve the video before any animation starts. A changed sentence after animation means new timing, new scenes and a new voice take.

Our explainer video script guide has templates and examples.

Time: 1 to 3 hours, more if several people need to agree.

Step 2: Turn the script into a timed scene list

A scene list is the bridge between words and prompts. Make a simple table: one row per scene, three to eight seconds each.

(Invented example.)

Three rules make the list useful:

• One visual idea per scene. If you write "and" in the visual column, split the scene.
• Show what the words can't. The visual should add the picture, not repeat the sentence as text.
• Note every exact string. Prices, percentages, product names and button labels go in the table, spelled exactly as they must appear.

Time: 1 to 2 hours. See also our guide to explainer video storyboards.

Step 3: Pick the generator by the output you need

Choose by the file you need at the end, not by the demo reel.

The deciding question: will someone read something on screen? Then use a tool that builds editable layers or code, so text and numbers stay exact. Generated video can garble words and can't be corrected one word at a time.

Step 4: Prompt one scene at a time

The most common mistake is one giant prompt for the whole video. You get something, it's 70% right, and every fix changes three other things. Prompt scene by scene instead.

A prompt structure that works

1. Format and length: "16:9, 5 seconds."
2. Layout: what sits where. "Large headline left, illustration of a laptop right."
3. Exact text: in quotes. "Headline: 'Approvals take one click.'"
4. Motion: verbs and order. "Headline fades up first, then the cursor moves to the button and taps; a check mark draws itself."
5. Timing and easing: "Each entrance 0.4 seconds, ease out; hold the final frame 1.5 seconds."
6. Style and brand: "Flat illustration, no gradients. Colors #1B2A4A, #F2694B, #F7F3EE. Font: Inter Bold for headlines."

Weak prompt: "Make a cool animation about invoice software."

Strong prompt: "16:9, 5 seconds. Flat illustration, brand colors #1B2A4A (background), #F2694B (accent), #F7F3EE (text). A paper invoice slides in from the left into an open laptop at center. On the laptop screen, three fields light up one after another with the labels 'Vendor', 'Amount', 'Due date'. Each element enters in 0.4 seconds with ease-out. Hold the final frame for 1.5 seconds. No other text."

Follow-up prompts: one change at a time. "Make the invoice slide slower, 0.8 seconds." "Swap the field order: Amount first." If you change five things at once and it gets worse, you won't know why.

For coding agents (Remotion, HyperFrames): the same structure works. Add "render scene 1 only and show me a still of the final frame" before you ask for the full video. A wrong scene caught at frame one costs a minute; caught after a full render, it costs much more.

Time: 15 to 45 minutes per scene for a beginner, much less once you have a template.

Step 5: Edit, don't regenerate

Once a scene is 90% right, stop prompting. Open the editor and fix the last 10% by hand: nudge a position, retype a word, adjust a delay.

• Lock finished scenes. Duplicate the project or save a version before experimenting.
• Build a scene template. When scene 3 looks right, reuse its layout, type sizes and timings for later scenes. Consistency is what makes a video feel professional.
• Keep the whole in view. Watch all scenes back to back every few changes. Scenes that work alone can clash in sequence: different speeds, different visual weight, colors that drift.

Time: 1 to 3 hours for a 60-second piece.

Step 6: Add voice-over, music and sound

• Voice-over: synthetic voices are now good enough for most business videos. Choose one voice and keep it for the whole series. Listen for names, product terms and numbers; fix pronunciation with spelling hints or by rephrasing. Our guide to explainer video voice-over covers AI versus professional voices.
• Retime to the voice. Generated scenes rarely match the recorded voice exactly. Adjust scene lengths so each visual lands on its words.
• Music: quiet, steady, without vocals under speech. Check the license covers commercial use and every platform you'll post on. More in music for explainer videos.
• Sound effects: sparingly. A soft click on a button press helps; a whoosh on every entrance gets tiring fast.

Time: 1 to 2 hours.

Step 7: Captions, formats and export

• Captions: many people watch muted, especially on social. Burn captions into social versions and add a caption file (SRT or VTT) for your website player and YouTube.
• Safe areas: on 9:16, keep text away from the top and bottom, where platform buttons and captions sit.
• Export settings: for most uses, 1080p, 30 frames per second, H.264 MP4. Use 60 fps only if the motion needs it, and 4K only if the platform shows it.
• One master, many cuts: export 16:9 first, then reframe for 9:16 and 1:1. If you planned the formats in the five decisions, this takes minutes, not hours.

Time: 30 minutes to 2 hours.

Step 8: The ten-minute quality check

AI tools don't check their own facts. Watch the final file once with this list:

1. Every word spelled correctly, including product and company names.
2. Every number, price and date matches the script table.
3. Brand colors and fonts are the same in every scene.
4. Each text stays on screen long enough to read twice (about one second per three to four words, plus half a second).
5. Text has enough contrast against its background.
6. Nothing flashes rapidly; fast flicker is uncomfortable and can be a risk for people with photosensitive epilepsy.
7. The voice matches the visuals; no scene runs ahead or lags.
8. Audio levels: voice clearly above music, no clipping.
9. The call to action is visible and readable at the end.
10. Watch it once on a phone, muted. Does it still make sense?

How long the whole workflow takes

Our rough estimates for a first-time user making a 60-second video, based on our own production experience:

The second video goes much faster: you keep the template, the brand kit and the prompt patterns. The first one is an investment in a skill.

When a generator is the right call, and when it isn't

A generator is a great fit when you need many short pieces: weekly social clips, feature announcements, data snippets, title cards. Learn one tool, build templates, and each new piece takes an hour or two.

It's a weaker fit when you need one important explainer: the video on your homepage, in your sales emails or at a trade show. There, story, clarity and consistency across scenes matter more than speed, and the 10 to 28 hours above fall on someone whose time is probably better spent elsewhere.

That's the work we do at Capri Video. We focus on nailing explainers so you don't have to. You paste your website link; we write the script, design the scenes in one of our styles, animate, add a professional voice and music, and our team reviews every draft before you see it. You get a watermarked draft within 24 hours and three revision rounds. To keep it, subscribe from $99 a month plus sales tax, which covers one video up to two minutes.

Frequently asked questions

What is an AI motion graphics generator?

A tool that creates animated text, shapes, icons, charts or whole scenes from a written prompt. The better ones produce editable layers or code, so you can still change words, colors and timing after the first result instead of generating everything again.

Is there a free AI motion graphics generator?

Several tools have free plans, including Jitter, Canva and Lottie Creator, and the code-based frameworks Remotion and HyperFrames are free to run under their own licenses. Free plans usually limit resolution, AI credits or add a watermark, so they're good for testing a workflow, less so for a published business video.

How do I write a good prompt for motion graphics?

Describe one scene at a time and include the exact on-screen text, the layout, how each element moves, how long the scene lasts, your brand colors as hex codes and your font. Then change one thing per follow-up prompt, so you can tell which instruction did what.

How long should text stay on screen?

Long enough to read it twice at a relaxed pace. As a rule of thumb, allow about one second per three to four words, plus half a second, and never less than about 1.5 seconds for a short line.

Can AI motion graphics replace a motion designer?

For short, simple pieces such as title cards, social loops and basic charts, AI tools often get you close enough. For a story across many scenes, consistent characters and a polished brand feel, an experienced designer or a specialized studio still saves time and gets a better result.

How long does it take to make a 60-second motion graphics video with AI?

For a first-time user, plan one to three working days including script, scene plan, voice, music, revisions and checks. The animation step itself can shrink to a few hours with a good generator; the rest shrinks much less.

Your time is the scarce resource

Use a generator for the many small pieces, and keep your hours for the work only you can do. For the one explainer that has to win customers, hand us the hours.

Start for free with your website link and get a watermarked draft within 24 hours. To keep it, subscribe from $99 a month.