Docs, Decks, and Web Pages Into Talking Videos in One Click—VibeKnow, the One I Tested and Kept

Let me say this first: this is a recommendation episode. A company building AI video reached out and asked whether I’d help promote them.

Back when this account had just started and had only a few hundred followers, plenty of sponsorship offers came in and I turned every one of them down. They were usually either API relay shops or products that just weren’t good enough for me to force myself to praise. Today’s is genuinely useful, which is why I took it—and what I negotiated was no extra fee for me, in exchange for credits for viewers.

What VibeKnow is: doc / deck / web page → a publishable video

It turns a document, a deck, or a web page into a short video you can publish as is, in one click. I don’t think it counts as a NotebookLM substitute anymore—it has gone past NotebookLM.

Open the site and you can upload a file, or just give it a link or a block of text. Uploads support a lot of formats: PDF, Word, PPT, or plain TXT and Markdown. Once your material is in, it auto-generates a prompt—telling the agent how many minutes the video should run and what core idea it has to get across.

Four modes

• Flexible creation: you can freely edit the visuals along the way, then generate the video from the frames you’ve locked in.
• One-click video: let the agent do its thing and get a video immediately.
• Deck narration: upload an existing deck and, without changing its content or layout, add spoken narration to every page and turn it straight into a video.
• Illustrated video: break a piece of content into frame-by-frame illustrations that act it out.

What hit me hardest: turning a “web tutorial nobody sees” into something shareable

That illustrated video mode—I think it’s fantastic.

Followers know I used to build web pages to explain my skill methods. But web pages have a problem: they don’t travel well on the internet, and they get no organic traffic. VibeKnow turns a web tutorial into a short video people will genuinely pass around, and it lays the material out in a clear, easy-to-follow order.

Lots of templates, and it can use my own voice

It offers plenty of templates: a hand-sketched style, a straight explainer style, and a government/business style—great for an industrial park or a business unit making a briefing-style video. I went with a “spiral notebook” look.

One more thing worth mentioning: you choose by hand how many frames the video is built from. Say I want richer content and a tighter fit between narration and visuals. I can have those 120 seconds generate 12 illustrations—a new frame every 10 seconds on average, which makes the whole video fuller and more watchable.

The voice is a highlight too: it has plenty of built-in voices, and you can create a new one. I uploaded my own voice, so the final video is narrated in my voice. Honestly, the result is far better than what I produce myself.

Cost: from ¥200-300 a video to about ¥10-20

It keeps costs down remarkably well. I used to post a lot of animated illustrated videos on WeChat Channels, all generated with Nano Banana and Seedance. Back then one video, two to three minutes, ran me north of ¥200—sometimes ¥300. On VibeKnow the same video costs about ¥10-20 a minute.

It also ships a full CLI integration: hand the docs to Claude Code and CC will take your material and generate the video you want, the way you normally work or however you specify. I barely work through web interfaces at all anymore.

Viewer perk: signup credits + daily login credits

What I told the product team was: don’t pay me extra—letting more users try it is worth more. So right now, viewers who go through the link on the blog get nearly ¥50 in credits, plus another ¥10 in credits for every daily login.

What does that money buy? The signup credits let you jump right in and generate a good-looking video immediately. The daily credits let you make another one every couple of days, which covers everyday use fine.

What it really solves: turning a company’s document assets into video

I think it solves a real problem for companies—especially education, which sits on enormous volumes of training material, and traditional industries, which have piled up documents for years. Those documents are genuinely good assets. Turn them into video and the people inside can actually learn from them, while outside partners come away with a much better sense of the company’s background and culture.

That’s it for today. See you next episode.

Source: EP0046_audio.mp3 · ASR model gemini-2.5-pro (chunked parallel) · full text of the original recording

[00:00] Today’s video comes with a perk for followers, especially for friends in education and training, or in traditional industries who want to do corporate marketing. A product came to Pinpin and asked whether I could help promote it. Honest truth: back when this account was just starting, when I only had a few hundred followers, plenty of sponsorship offers came in, and I turned down every single one. Usually they were things like third-party relays, or products that were just badly made, the kind I couldn’t force myself to hype up. Today’s product is genuinely good. It can turn a document, a deck, or a web page into a publishable video in one click.

[00:26] I don’t even think it counts as a NotebookLM alternative anymore—it’s already a product that’s gone past NotebookLM. Open the product’s site and you can see you can upload a file, or hand it a link or a piece of text. For uploads it supports a lot of formats: PDF, Word, PPT, or plain TXT and Markdown files. Once we upload the file it automatically generates a prompt—telling the agent how many minutes of video it needs to make,

[00:51] and what the core idea it wants to get across is. There are actually four modes in here. You can see the first one is flexible creation: as you go you can freely edit the frames, then generate the video from the frames you’ve settled on. The second, one-click film, just lets the agent run with it and produce that content for you right away. Deck narration is when we’ve already uploaded an existing PPT—without changing its content or layout, it adds voiced narration to each page and turns it straight into a video.

[01:17] And then the illustrated video—this one, I think, is unbelievably useful. Followers know I’ve been building web pages to explain the methods behind some of my skills. The problem with web pages is that they don’t spread that well on the internet; they get no organic traffic. With VibeKnow, you can turn a web-page tutorial into a short video that really travels. And the way it puts things across is really well organized, really clean and easy to follow.

[01:42] It offers a lot of templates for us to choose from. Some are a simple line-drawing style, some are a really explainer-ish style, and there’s a government and business style that’s a great fit for our industrial parks or public-sector departments to use for report-style videos. I went with a ring-bound notebook look. One more thing worth mentioning: it lets us manually choose how many pages go into the generated video. So for example, to make the content in there really rich, with a good fit between the images and the text—

[02:08] for these 120 seconds, I can choose to generate 12 illustrations to carry it. That way there’s a frame roughly every 10 seconds, which makes the whole video feel fuller and more watchable. And the part I really love in here: it has a lot of built-in voices, and you can also create a new voice. I uploaded my own voice, so the video it finally generated is narrated in my voice. Let’s look at how this video came out. Honestly I think it’s far better than what I make—

[02:34] myself. Have you ever run into this? You work hard on a web report, hit Ctrl+P, save as PDF, and the result is a disaster. These are the browser’s three cardinal sins. One: the content gets chopped in half. It cuts hard at A4 height, so a card or a table often gets sliced in two—half on one page, half on the next. Two: the desktop layout collapses into the mobile layout. When printing, the browser re-evaluates breakpoints against the paper width, and a three-column grid quietly collapses to one column.

[02:59] Three: backgrounds and texture are all lost. So the result is stunning. And it keeps the whole cost well under control. The most hands-off part? These days I almost never work through a web UI at all. So it gives you a full CLI integration path. Just hand the docs to Claude Code and CC can go straight at it—we send it our material, and then working the way we’re used to, or just telling it what we want, it can generate the—

[03:25] video we want right away. And then last, the follower perk. What I told the product team was: don’t pay me anything extra. For a product this good, better that users get to try more of it. So right now, if followers go through the link on the blog to use it, they get close to 50 yuan in credit, and there’s another 10 yuan of credit for logging in each day. What does that money mean? Well, you can see I’ve had a lot of animated videos on WeChat Channels before—those were all made with Nano Banana and Seedance

[03:50] to generate the animation, that kind of illustrated video. Back then one video cost me around 200, even 300 yuan a piece, for two to three minutes of content. On the VibeKnow platform, generating the same video runs you a little over ten yuan a minute. With the credit they’ve given, you can start experimenting right away—generate a really nice-looking video immediately. And the daily credit means you can make one every couple of days, which is plenty for everyday use.

[04:15] I think it solves a big headache for companies, especially industries like education, with their mountains of training material, or traditional industries, which have piles of documents built up over the years. Those documents are genuinely great assets. Turning those assets into video lets internal people learn from them properly, and lets outside partners get a better sense of our company’s background or culture. That’s it for today. See you next episode, bye-bye.