AI Explainer Video Statistics 2026: What 2,300 Videos Made From Documents Show

By , founder and CEO of DistilBook · Published September 27, 2026

Ajith builds AI tools for explainer and motion-graphics video: DistilBook, and Framenet AI before it.

AI Explainer Video Statistics 2026: What 2,300 Videos Made From Documents Show

Since April 2026, more than 1,400 people have made over 2,800 videos with DistilBook. This page reports what the 2,308 finished videos show: what people turn into video, how long the videos run, which styles and languages they choose, and how long the whole process takes. Every number is an aggregate from our own production data.

What are the key numbers?

Across 2,308 finished videos made between 7 April and 26 September 2026, 86% started from a PDF, the median PDF was 5 pages, the median video ran 2.9 minutes, 83% used the hand-drawn whiteboard style, and the median video was finished 12 minutes after upload.

MeasureResultSample
Videos started from a PDF86%2,308 finished videos
Median PDF length5 pages (90% were 11 pages or fewer)1,984 PDF videos
Median video length2.9 minutes (90% were 12.1 minutes or less)2,306 videos
Median slides per video2 (90% had 6 or fewer)2,307 videos
Hand-drawn whiteboard style83% (motion graphics 17%)2,308 videos
Median time from upload to finished video12.3 minutes, including script review2,306 videos
Narrated in a language other than English42%371 recent videos

What do people turn into explainer videos?

Mostly PDFs. 86% of finished videos started from an uploaded PDF, 9% from a web link, 3% from images and 2% from a product page. The PDFs were short: the median was 5 pages and 90% were 11 pages or fewer, so people convert a lecture, a chapter section or a guide, not a whole book.

The short median matters for anyone planning educational video. People do not upload a 300-page textbook and ask for one video. They pick the pages for one topic and make one video about it, which is also what the research on video length recommends.

Starting pointShare of finished videos
PDF upload86.0%
Web link (article or page)9.3%
Images3.2%
Product page1.6%

How long are AI explainer videos made from documents?

Short. The median finished video ran 2.9 minutes (173 seconds) and 90% ran 12.1 minutes or less. The median video had 2 slides and 90% had 6 or fewer. The average is higher, 5 minutes, because a small number of long videos pull it up.

That lines up with the best evidence on attention: across 6.9 million edX video sessions, the median viewing time was at most 6 minutes whatever the video's length. Most people making explainers from documents are already inside that limit.

Do people choose whiteboard or motion graphics?

Whiteboard, by a wide margin. 83% of finished videos used the hand-drawn whiteboard style and 17% used motion graphics. Among motion videos with a recorded style, the hand-drawn education style was picked 40% of the time, cut-paper collage 12% and isometric illustration 9%.

A drawn explanation that appears as the narration reaches it is what most people want from an explainer. In the same edX study, tutorials drawn on a tablet were watched 1.5 to 2 times longer than slide-based videos.

Motion style (137 videos with a recorded style)Share
Hand-drawn education40.1%
Cut-paper collage12.4%
Isometric illustration8.8%
White and navy line art5.1%
All other styles33.6%

How long does it take to turn a document into a video?

The median video was finished 12.3 minutes after the document was uploaded, and a quarter were finished within 8.8 minutes. That time includes the pause where the creator reads and edits the narration script before any artwork or audio is generated.

Longer videos take longer. Measured on the same finished videos, from upload to the finished video:

Slides in the videoMedian minutes from upload to finishedVideos
19.61,027
210.3252
3 to 412.9615
5 to 717.5192
8 to 1030.3178
11 or more50.741

Which languages are explainer videos narrated in?

English is the largest, but 42% of recent videos were narrated in another language. In 371 recent videos with a recorded narration language, 58% were English, 13% Russian, 10% Turkish, 7% Portuguese, 4% Ukrainian, 3% Spanish, 2% Hindi, 2% French and 1% German.

Explainers are increasingly made in the learner's own language rather than translated afterwards. DistilBook narrates in 19 languages, and the script is written in the chosen language from the start.

Narration languageShare of 371 recent videos
English58.0%
Russian13.2%
Turkish10.2%
Portuguese6.7%
Ukrainian4.3%
Spanish3.0%
Hindi1.9%
French1.9%
German0.8%

Who are these videos made for?

Mostly learners. Where creators described their audience in words we could group (136 recent videos), 51% named students or learners, 23% a general or beginner audience, 14% employees or trainees, 8% professionals and 4% customers.

This is a keyword grouping of a free-text field, so treat it as direction, not precision. It matches what we see elsewhere: teachers, course creators and people preparing for exams make most explainer videos from documents.

How was this data collected?

From DistilBook's production database: every video that finished between 7 April and 26 September 2026 (2,308 videos). We counted settings and file metadata only, such as file type, page count, length, style and language. No document content or customer details were read or published.

The headline figure of more than 1,400 people and over 2,800 videos also counts videos that were generated on the free plan but not unlocked. Language, motion style and audience are only recorded by our newer create forms, so those sections give their own, smaller sample sizes. You are welcome to cite these numbers with a link to this page.

  • Source: DistilBook production data, finished videos, 7 Apr to 26 Sep 2026.
  • Video length research: Guo, P. J., Kim, J., and Rubin, R. (2014). How video production affects student engagement: an empirical study of MOOC videos. ACM Learning at Scale.