How to make a whiteboard animation, step by step

Whiteboard animation looks like a drawing problem. It isn't. It's a writing problem with drawings attached. Get the script right and even simple visuals land. Get it wrong and no amount of slick draw-on animation saves it.

Here's the whole process, in order, with the honest version of each step.

Step 1: Write the script first

Not the visuals. Not the tool choice. The script.

A whiteboard video is a narrated explanation, so the narration is the skeleton of everything else. Write it like you'd explain the topic to one smart friend at a whiteboard. Short sentences. One idea at a time. Say it out loud as you write, because text that reads fine often sounds stiff when spoken.

Two practical checks before you move on. First, length: people comfortably watch 60 to 120 seconds of explainer before attention slips, and that's roughly 150 to 300 words. Run a word count on your draft and see where you stand. If you're at 800 words, you don't have a video, you have three videos.

Second, cleanliness: if your draft came out of a doc or a chat with an AI, it's probably full of markdown, bullets, and formatting that will trip up your narration. Strip it down to plain spoken text before you record.

Step 2: Plan the scenes

Read the script and mark the natural beats, the points where the idea shifts. Each beat becomes a scene with one visual.

The discipline here is one drawing per idea. When a scene tries to illustrate three things at once, viewers stop knowing where to look. A crude sketch of the right concept beats a beautiful drawing of a vague one.

You don't need storyboard software. A numbered list works: beat 1, stick figure at a fork in the road; beat 2, two arrows diverging; and so on. What matters is that every sentence of narration has a visual answer to "what is the hand drawing while I say this?"

Step 3: Produce it (three routes)

Now, and only now, pick your production route. There are three.

Route A: draw it for real. Film a hand drawing on paper or a whiteboard, or record a tablet screen, then speed up the footage and cut it to the narration. This is the most authentic look and the most work. Budget hours per finished minute, and expect the sync between drawing and voice to take several passes. Worth it when the handmade feel is the point.

Route B: use an editor. Tools like Doodly or VideoScribe fake the drawing hand for you. You pick images from their library, place them in order, set durations, and the software animates the sketch-in. This is the middle path: no drawing skill needed, but you're doing all the assembly. The common trap is the library not covering your topic, so check that before committing to a tool. Plan an afternoon for your first video and expect the timing step (nudging scenes so they land with the voiceover) to eat most of it.

Route C: generate it. AI tools take the topic or the script and produce the whole video for you: a finished, narrated whiteboard animation. Rekka works this way. The tradeoff is control: you're reviewing and regenerating rather than placing every element by hand. For straightforward explainers this route turns the whole process above into minutes.

Note what stays the same across all three routes: the script and scene thinking. That work is never wasted, whichever way you produce.

Step 4: Record the voiceover (Routes A and B)

If you're narrating yourself, a few things make an outsized difference.

Record standing or sitting upright, and smile slightly while reading. It genuinely changes the sound. Read at a natural pace, around 150 words per minute; rushing is the most common amateur tell. Do three full takes rather than fixing sentences one by one, because patched-together audio sounds patched together. And use whatever quiet room you have with a decent USB mic. Room echo hurts more than mic quality.

If recording yourself is the step you dread, that's a signal Route C exists for.

Step 5: Sync, render, and check

Assemble scenes against the narration and watch the whole thing once as a viewer, not a creator. You're checking three things: does each drawing appear when the words reference it, is any scene lingering after its point is made, and does the pacing sag anywhere. Trim ruthlessly. Then render at 1080p and watch the export too, since some tools produce artifacts that only show in the final file.

The mistakes that actually hurt

After you've seen enough whiteboard videos, the failures repeat:

• Scripts written for the eye, not the ear. If it sounds like a blog post being read aloud, rewrite it.
• Too long. The second minute has to earn itself. Most don't.
• Crowded scenes. One idea, one drawing.
• Mismatched visuals. Generic clip art of "businessman pointing at chart" while the narration explains DNS. Viewers notice, then trust drops.
• Skipping the out-loud read. Every stiff sentence in a finished video survived because nobody said it out loud first.

None of these are tool problems. They're all fixable in the script and the scene plan, before any software opens. Which is the point: make the thinking good, and the production route (hand, editor, or AI) is just a question of how much of your weekend you want back.