Blog post to AI video: cut an article into a 50-second shot list with one presenter

What the blog-to-video tools do

Search for how to turn a blog post into a video and page 1 is tools that take a URL. Paste the link, the tool summarises the article, an avatar or a stock-footage montage reads the summary, and captions go on top. The advice around them is sound as far as it goes: cut the article to a script of 100 to 150 words, export 9:16 at 1080 by 1920, stay under a minute, and do not dump the whole post into the script.

What none of them hands you is the step between the article and the video: which paragraphs become which shots, how long each shot is, how many words fit in it, and what the whole thing costs to render. That step is where an article-shaped video turns into a watchable one. This page is that step, the way we do it for our own posts.

The budget comes first: 50 seconds is about 110 words

Narration and on-camera lines both run at about 2.5 words a second at a brisk pace. Leave half a second of air at the start and end of each shot so no line starts on the first frame, and a 50-second video holds roughly 110 spoken words.

A typical blog post is 1,200 to 2,000 words. So the job is not summarising, it is choosing: about 7% of the article survives. Decide that number before you open the post, or the script will be 300 words and the video will be two minutes long.

What to cut from the article

Most of a blog post is there for reading, not for watching. We cut these first:

• The introduction. Posts warm up for two paragraphs. Video cannot. The hook replaces it.
• The search-engine scaffolding. Definitions, "what is X", repeated keywords. Nobody scrolling needs them.
• Long lists. "11 ways to" becomes three ways, the three you would defend.
• Caveats and footnotes. Keep the one that changes the advice. Drop the rest.
• The conclusion. It restates the post. The video ends on the last point and the call to action.

What stays is the headline claim, three to five points that carry it, and one concrete example or number per point.

Paragraph to shot: the map we use

Each surviving piece of the article becomes one shot with one job and a word cap.

shot	comes from	length	words, brisk
hook	the headline, turned into a claim or a question	4 s	about 7
point 1	first section's main sentence	8 s	about 17
point 2	second section's main sentence plus its number	8 s	about 17
point 3	third section, its example	8 s	about 17
point 4	fourth section, or the strongest counter-point	8 s	about 17
point 5	the takeaway the post builds to	8 s	about 17
close	what to do next	6 s	about 12
total		50 s	about 104

If a point needs more than 17 words, it is two points, or it is the wrong point. Rewriting it to fit is faster than rendering a 12-second shot nobody watches to the end. Our words per shot guide has the table for every length.

One presenter, not an avatar

The tools on page 1 give you a stock avatar reading the summary. We use a presenter character instead: a person or a 3D animated figure made once from a character reference sheet, then attached to every scene. The sheet is 30 credits, made once, and the same face hosts every post you turn into a video after that. That is what makes a series of repurposed posts look like one channel instead of five unrelated clips. The reference sheet guide covers making one that holds.

The presenter does not need to be on screen every shot. In our builds they speak the hook, two or three of the points and the close, on camera. The other points play as silent shots of the thing being talked about, with the presenter's voice over them.

The render plan and the credit total

On Kling O3 Pro at 720p, with the presenter attached (Kling takes up to 5 characters in a scene and charges nothing extra for them):

shot	what is on screen	audio	credits
hook, 4 s	presenter, to camera	voiced	120
point 1, 8 s	presenter, to camera	voiced	240
point 2, 8 s	the example, no presenter	silent	192
point 3, 8 s	presenter, walking and talking	voiced	240
point 4, 8 s	the example, no presenter	silent	192
point 5, 8 s	presenter, to camera	voiced	240
close, 6 s	presenter, to camera	voiced	180
total, 50 s			1,404

With our retake allowance (one and a half times each shot), budget about 2,100 credits for the first post. The narration over the two silent shots is recorded in your audio tool and laid under the cut. If you would rather the presenter voice every shot, render all seven voiced: 1,500 credits, 96 more.

On the two silent example shots, Veo 3.1 Fast is cheaper still: 8 seconds silent is 160 credits instead of 192, as long as no character is attached.

One post, four videos

The same shot list gives you more than one video. Each point shot is a self-contained 8-second clip with its own line, so:

• The main cut, 50 seconds, for the post itself.
• Three cutdowns, each the hook plus one point plus the close, about 18 seconds.

The cutdowns cost no extra renders, only exports: 20 credits each, free from the Creator plan up. Captions are 8 credits per started minute, also free from Creator up. One article becomes four uploads for the price of one render plan.

The order we work in

1. Count the budget: 50 seconds, about 110 words.
2. Strip the post: intro, scaffolding, long lists, conclusion.
3. Pick the headline claim and three to five points, one number or example each.
4. Fill the shot map and cut every line to its word cap.
5. Mark which shots the presenter speaks on camera; those render voiced.
6. Render, retake the misses, cut the main video and the three short ones.

We build the shot list in the studio straight from the stripped draft, one scene per row, and the total under the render button tells us the cost of the post before any of it is billed. If you already have a script rather than an article, the script to video guide starts one step later.

FAQ

How do I turn a blog post into a short AI video?

Cut the post to about 110 words for a 50-second video: a hook, three to five points with one example each, and a close. On ReelDouble each piece becomes one scene of 4 to 8 seconds with the same presenter character attached.

How many words of a blog post fit in a 60 second video?

About 2.5 words a second at a brisk pace, minus half a second of air per shot, so roughly 130 words for 60 seconds and 110 for 50. Most of a 1,500-word article has to go.

How much does it cost to turn an article into an AI video?

A 50-second, seven-shot video on Kling O3 Pro on ReelDouble costs about 1,404 credits with five voiced shots and two silent ones, or 1,500 fully voiced. Budget one and a half times that for retakes.

Can I use the same presenter in every blog video?

Yes. On ReelDouble you make one character reference sheet for 30 credits and attach it to every scene, so the same presenter hosts every post you turn into a video.

Should a blog post video be one long video or several short ones?

Both, from one render. Each 8-second point shot stands alone, so on ReelDouble one 50-second build also cuts into three 18-second shorts that cost only an export each, free from the Creator plan up.