How to Make a Video Abstract

A video abstract is a short video that sums up a research paper for people who have not read it. The publishers we checked allow two to four minutes. It covers four things: the question, the method, the finding, and why the finding matters. You can film one yourself or draft one from the paper PDF with a video abstract generator.

Key takeaways

Question	Short answer
What is a video abstract?	A short video summary of a research paper, for people who have not read it
How long should it be?	Two to four minutes. Check your journal’s own page
What goes in it?	The question, the method, the finding, and why it matters
What file do journals want?	Usually a 16:9 MP4. Two of the three we checked also want a transcript
Can I draft one from the PDF?	Yes. Upload the paper and use Summarize, then check every claim yourself
Do journals accept AI-made ones?	The three author pages we read don’t say. Ask your editor first

What journals ask for

Every publisher sets its own rules, so read your journal’s author page before you make anything. We read three publishers’ pages on 11 October 2026. Their limits differ more than you’d expect.

	Taylor & Francis	BMJ	IOP Publishing
Length	2 minutes 20 seconds or less is “optimal”	No more than 4 minutes	No longer than 4 minutes
Shape	16:9	16:9 only	16:9 or 4:3
File type	.mov, .mpg, or .mp4	.mov, .mpg, .avi, or .mp4 (.mp4 preferred)	.mov, .mpg, or .mp4
Smallest frame	720p	720 x 576 pixels	320 x 240 pixels
Frame rate	24, 25, or 30 fps	25 fps	25 to 30 fps
File size cap	Not stated	Not stated	100 MB
Transcript	Required, with the visuals described	Not stated on the pages we read	Required, as a .txt file
When to send it	After acceptance. Medicine and health journals want a storyboard at submission	After acceptance only	Before or after acceptance

Sources: Taylor & Francis Author Services, the BMJ Author Hub, and IOP Publishing Support.

Three rules in those pages catch people out.

1. IOP bans music. Its guidelines say a video abstract “must not include music soundtracks,” and that covers open source music too.
2. BMJ won’t take one early. It asks authors not to upload a video abstract with the first submission. It invites one after the paper is accepted, and it says publication is not guaranteed.
3. A transcript is part of the job. Taylor & Francis and IOP both require one so the video is accessible.

These pages cover three publishers. Your journal may ask for something else, or for nothing. Many journals don’t mention video abstracts at all.

What the pages say about AI

None of the three pages mentions AI tools. That silence is not a yes. All three describe the author on camera or speaking. Taylor & Francis asks you to introduce the work “in your own words.” IOP “strongly” encourages footage of yourself with other material.

So treat an AI-narrated video as a draft or as a version for your own channels, unless your editor says otherwise. Ask before you submit one. Check the journal’s AI policy too, and disclose what you used.

The four-part script

A video abstract answers four questions in order. Taylor & Francis frames it almost the same way: what did you want to answer, how did you go about it, and what did you conclude.

1. The question

Open with the question the paper asks, in words a stranger would use. Skip the literature review. Two sentences is enough: what nobody knew, and why you wanted to know it.

2. The method

Say what you did in one breath. Name the sample, the setting, and the test. Leave the statistics in the paper. A viewer needs to trust the finding, and a plain account of the method earns that trust.

3. The finding

State the result first and the detail second. This part gets the most time. Include the main limit or exception here. Don’t save it for the end, because a finding with its caveat attached is harder to misquote.

4. Why it matters

Close with what changes because of this result. Who should care, and what should they do or study next? IOP’s guidelines ask for the same thing: discuss what could develop in the field as a result of the work.

Papers build up to the result, and video can’t wait that long. We’d give the question twenty seconds at most and get to the finding before the one-minute mark.

Worked example from an open-access paper

Our example is a paper that anyone can read and reuse. “Domestic cats (Felis catus) discriminate their names from other words,” by Atsuko Saito, Kazutaka Shinozuka, Yuki Ito, and Toshikazu Hasegawa, appeared in Scientific Reports in 2019. It is open access under a CC BY 4.0 license.

We wrote the script lines below ourselves, from the paper’s abstract, to show the four parts. We did not run the paper through AITuber, and the authors did not review our wording.

Part	Example line
Question	”Dogs get most of the research. Can a cat tell its own name from other words?”
Method	”Researchers played each cat four other words in a row, then its own name. They watched to see if the cat reacted afresh.”
Finding	”Household cats that had stopped reacting to the other words responded again to their own names. It worked even when a stranger spoke.”
The caveat	”Cats living in a cat café could tell their names from ordinary nouns. They could not tell them from the names of the other café cats.”
Why it matters	”The authors conclude that cats can tell human words apart by their sounds.”

That is about 100 spoken words, or under a minute. A two-minute version has room for one more sentence on the method and one on what to study next.

Drafting it from the PDF

The same script can start as a draft from the paper itself.

1. Cut the reference list and save a new PDF. A long list of citations pulls a script toward names and dates. Keep the file under the 25 MB upload limit.
2. Upload the PDF. AITuber reads the text layer, and it runs OCR on a scanned paper. The upload takes PDF only, so export a Word manuscript to PDF first.
3. Pick Summarize. It reads the full paper and writes around the main points. Read it out would start at the title page and run out of time.
4. Set the length to your journal’s number, and choose 16:9.
5. Add a one-line steer. This line shapes the script more than any other setting.

For the cat paper we’d write this steer:

Open with the question. Explain it for a general audience. Keep the cat café exception. Claim nothing the paper does not claim.

Two limits apply. First, AITuber makes its own visuals for each scene. It does not lift the figures from your PDF, so Figure 3 will not appear unless you add it afterward in an editor. The result is a video about the ideas in the paper, and your pages never show on screen.

Second, the narration is an AI voice. If your journal wants you on camera, the draft is still useful. Read the finished script to the camera yourself, and use the visuals as cutaways.

Our guide on turning a PDF into a video covers the upload steps in more detail.

Review checklist before you submit

The author stays responsible for every claim. A summary can round a number, drop a caveat, or state a result more strongly than the paper does. Read the script next to the paper before anyone else sees the video.

1. Check each number against the paper. That includes sample sizes, effect sizes, and dates.
2. Check the verbs. The cat paper says cats “discriminate” their names. A script that says cats “understand” their names claims more than the study showed.
3. Keep the limits in. If the abstract names an exception, the video should too.
4. Look for corrections. The cat paper has an author correction dated 10 September 2019. Make sure your script matches the current version.
5. Listen to the hard words. Species names and compounds can come out wrong in a spoken voice. Fix the spelling in the script until they sound right.
6. Get every co-author to watch it. Their names are on the work.
7. Match the spec. Check the length, shape, file type, frame rate, and file size against the journal’s page. Export at 1080p if there is a size cap.
8. Remove music if the journal bans it. IOP does.
9. Write the transcript. Copy the final script into a text file. Taylor & Francis also wants the visuals described, so add a line for each scene.
10. Tell the editor how you made it.

We’d never skip step two. Overclaiming is the easiest mistake to make in a short script, and it is the one a journalist will quote.

Where to post it

The journal page is the first home. A publisher that accepts your video hosts it next to the article.

Your rights to post it elsewhere depend on the journal. IOP says its copyright lets authors host the video on their own website. BMJ publishes video abstracts under the same copyright terms as the article. Read your agreement before you upload anywhere else.

Where you have the right, three more places are worth the effort.

1. Your lab or department site. Put it at the top of the publication’s page, above the written abstract.
2. YouTube. Use the paper’s title in the video title, and put the DOI link in the first line of the description.
3. LinkedIn. Post it with one line on why the finding matters, and tag your co-authors.

A vertical cut helps on social feeds. Journals want 16:9, but a 9:16 version of the finding alone suits Shorts and Reels. For a lay audience, an AI explainer video made from a plain-language script may travel further than the formal abstract. The PDF to video page handles reports and other documents the same way.

Real examples

Science explainers made with AITuber

Narrated science videos from the public AITuber catalog. A video abstract uses the same script, voice, visuals, and captions.

Frequently Asked Questions

Basics

What is a video abstract?

A video abstract is a short video that sums up a research paper for people who have not read it. It covers the question, the method, the finding, and why it matters.

How long should a video abstract be?

Two to four minutes suits most journals. BMJ and IOP Publishing cap it at four minutes. Taylor & Francis calls 2 minutes 20 seconds or less optimal. Check your own journal’s page.

Do journals require a video abstract?

Most don’t. Some invite one, and many don’t mention it. BMJ asks authors to wait until the paper is accepted before sending one.

Making one

What should a video abstract include?

Four parts in order: the question the paper asks, how you tested it, what you found, and why it matters. Give the finding the most time.

Can I make a video abstract from my paper’s PDF?

Yes. Upload the PDF to AITuber, pick Summarize, set the length, and add a one-line steer. Then check every claim in the script against the paper.

Will it show the figures from my paper?

No. AITuber makes its own visuals and does not extract figures from the PDF. Add a figure afterward in an editor if it has to appear.

What file format do journals want?

Usually an MP4 in 16:9. Taylor & Francis, BMJ, and IOP Publishing all accept .mp4. IOP sets a 100 MB cap.

Rules and cost

Do journals accept AI-generated video abstracts?

It depends on the journal, and many pages don’t say. The three author pages we read don’t mention AI tools. Ask your editor before you submit one, and disclose what you used.

Who is responsible for what the video says?

The authors are. Read the script next to the paper, check each number, and keep the limits in.

What does it cost to draft one?

Signup on AITuber includes 100 free credits with no card. A 30-second Short with AI images costs about 35 credits, and a two-minute abstract costs more. The AITuber Creator plan is $39 a month for 2,500 credits.

---

How we researched this: we read the Taylor & Francis Author Services page on video abstracts, the BMJ Author Hub pages on video abstracts (overview, top tips, and technical specs), the IOP Publishing Support page on video abstracts, and Wiley’s author page on video on 11 October 2026. Publisher rules change, so read your journal’s own page before you make anything. The worked example is Saito, A., Shinozuka, K., Ito, Y. and Hasegawa, T., “Domestic cats (Felis catus) discriminate their names from other words,” Scientific Reports 9, 5394 (2019), doi 10.1038/s41598-019-40616-4, used under CC BY 4.0. We wrote the example script lines from its abstract, and the authors have not reviewed them. The upload steps follow AITuber’s PDF form as it stands on that date. AITuber prices and credits come from our live plan list. AITuber is our product, so we’ve said where it falls short of what a journal asks for.

The video abstract generator page opens the upload form with Summarize already picked.