explain.ink
Hand-drawn explainer videos from any topic
Every "AI video" tool I tried does the same thing: it takes your script and pairs it with stock footage. That works until the topic is even slightly specific. Explain how a heat pump moves heat, and you get a montage of people pointing at radiators.
So explain.ink draws instead. It writes the story, splits it into scenes, and then actually draws each scene stroke by stroke in one of 18 styles, with narration recorded to match. You get a 16:9 MP4 you can use commercially.
Drawing turned out to be the easy part. The hard part was making the story worth drawing: early versions produced technically correct scenes that explained nothing, because the script and the picture were generated separately and never agreed with each other. Tying them together took longer than the renderer did.
If you make explainers for your own product, I'd like to know what you do about this. Do you script first and find visuals after, or the other way round?
3 Comments
The fact that the story and visuals initially drifted apart is a pretty interesting problem. Curious what the finished explainers look like once the two are tied together.
The drift came from writing the script and picking the picture in two separate passes. Now one pass returns the narration, the caption and the drawing for every scene, plus a small visual vocabulary fixed once for the whole video, so the sentence you hear and the thing being drawn come out of the same plan. Fair warning if you go looking: the gallery on the site is older than that change, so it shows the before, not the after. I owe it a refresh.
That sounds like a much cleaner production model. The shared visual vocabulary is especially interesting — I’d be curious to see how much better the finished explainers look after that change.