D
Dynamo.sh open-source whiteboard video
Start free

The best whiteboard animation software in 2026 — and the free, open-source alternative

Whiteboard videos work because a hand drawing an idea is easier to follow than a slide of it. The tools that make them are mostly good. They are also mostly subscriptions that meter your output, keep your source material, and hand back a file you cannot open. This compares seven of them honestly — including where they beat us — and explains what a free, self-hostable one changes.

Written by Appau Ernest Kofi Mensah · Updated September 2026

The short version

ToolTypical costMade forSource availableAPI
VideoScribe~$18–35/monthHand-drawn marketing and trainingNoNo
DoodlyOne-off, several hundredDoodle-style sales videosNoNo
Renderforest~$10–50/monthTemplated marketing videoNoPartial
Animaker~$20–80/monthCharacter animation, explainerNoNo
Vyond~$50–100+/monthCorporate L&D at scaleNoEnterprise
Powtoon~$20–90/monthPresentations and explainerNoNo
Explee~$15–40/monthWhiteboard, educationNoNo
Dynamo.shFree · self-hostExplainers, teaching, STEM workingYes, MITYes, open

Prices move and tiers change; check them before you buy. The columns that do not move are the last two, and they are the ones this article is really about.

What the paid tools are genuinely good at

It would be dishonest to write this without saying it plainly. The commercial tools have libraries. VideoScribe and Doodly ship thousands of hand-drawn assets, drawn by illustrators, in a consistent style. Vyond has character rigs with lip-sync and a scene editor that a marketing team can use on their first afternoon. Animaker and Powtoon have templates that get a decent video out of somebody who has never made one.

If your job is to produce a polished sixty-second brand video this week and you are not technical, buy one of them. That is what they are for, they are good at it, and no amount of open source substitutes for an illustrated asset library and a support line.

Where they cost you

The meter

Almost every one prices by output: minutes rendered, videos exported, seats. That is fine until the thing works, at which point wanting more of it becomes expensive precisely because it is working. A teacher making one video per lesson, or a support team making one per feature, hits this quickly.

The export

You get an mp4. You do not get the thing that produced it. Change a number in a formula six months later and you are back in their editor, on their plan, hoping the project still opens. There is no format you can keep, diff, or regenerate from.

The upload

Whiteboard tools are used for training material, internal process, unreleased product, exam content and patient information. All of it goes to somebody else's servers to be processed. For a lot of organisations that is the end of the conversation, and it is a reasonable place to end it.

What Dynamo.sh does differently

It is an engine rather than an editor. You give it a prompt, a PDF or your own recorded voice; a language model writes the narration and decides what belongs on each board; a deterministic renderer draws it with a hand, in a browser, frame by frame, straight into ffmpeg.

The timeline is a file

Every video is a JSON document: every element, its position, its colour, the exact moment its stroke begins. You can read it, edit it by hand, keep it in git, diff two versions, and render it again in a few seconds without calling a model at all. That is the piece the hosted tools cannot give you, because it is the piece they sell.

Ink is timed to speech, not to a stopwatch

Each element is anchored to a phrase in the narration, and the renderer places the stroke so that it finishes just before the word is said. A drawing that completes exactly on the word reads as late; one that lands a beat early reads as synchronised. It is a small rule that accounts for most of the difference between a whiteboard video that feels hand-made and one that feels generated.

It is built for working, not only for marketing

Most whiteboard tools assume a sales script. This one was pushed hard in the other direction: numbered steps, formulas that stack down the board one line at a time, a diagram given its own panel so the working can run beside it, a camera that comes in to read a full board and pulls back out. If what you make is a worked solution rather than a pitch, that matters more than an asset library.

Streaming

There is an API that streams the timeline scene by scene as it is planned, so a learner can start watching the first board while the rest is still being made, instead of staring at a progress bar for four minutes. The receiver does the drawing, so it is a few kilobytes of JSON rather than video.

Where we are worse, and you should know

What it actually costs to run

There is no per-video price because there is no per-video cost beyond a machine. The real cost of a video is the model that writes it and the voice that reads it, and those are billed by the providers to whoever owns the key. So the line is drawn exactly there: your first video is on our keys, and after that you add your own — a Groq key and a Fish Audio key, both free to create, a few minutes each, and pennies for a short explainer. Self-host and there is no limit at all, because then the machine is yours too.

So which should you pick?

The honest summary: if you want an illustrated asset library and a support line, pay for one. If you want an engine you can read, script, self-host and keep, that is what this is — and it is free either way.

Make one free and judge for yourself Read the source