How an article is made
From a link to a finished page usually takes a few minutes. You can close the tab — the work won't stop.
1. The words
First come the words of the video with their timecodes: captions made by the author, or YouTube's automatic ones. If there are none, the sound is downloaded and the speech is recognized — otherwise a video where people talk over engine noise would lose its words.
2. Who is speaking
Before writing, the transcript is read once to answer a few questions: how many people speak, whether one of them leads, what kind of video it is — a monologue, a channel hosted by two, an interview, a conversation, a song — and where the sponsor reads are.
That decides whose voice the article is in. One narrator writes as “I”. Hosts write as “we”, and what's personal stays with whoever said it. An interview is written by its host, and the guest speaks in their own words. The third person is left for videos where nobody from the channel speaks. Ads are cut out of the transcript before a single word is written.
3. The text
The article is written from the transcript, in the language of the video. Whatever only makes sense on screen goes away: “as you can see here” turns into what could be seen, and “link in the description” disappears. Facts worth checking get their timecodes, and pictures go where they add something the words don't.
A long video is written in parts, one after another, and each part sees the headings of the ones before it. The reader shouldn't be able to tell where one part ends and the next begins.
4. The frames
Frames are taken at the seconds the text points to. Several frames around each moment are tried and the sharpest one wins, so a black screen or a blurred transition doesn't slip in. A frame is never stretched: vertical footage stays vertical.
Then every frame is looked at once more: captions are corrected to what is really in the picture, and empty or repeated frames are dropped.
Videos without words
A hike set to music or a drive without commentary has no article to write — only one to invent. So the density of speech is measured before anything is written. When there are almost no words, Storiteca makes a photo report instead: frames at scene changes and a short text built only from what is really there — the video description, the author's chapters, the signs and captions visible in the frames. No one is handed an “I” for words they never said.
What doesn't get a page
Videos shorter than a minute: watching one is faster than reading about it. Private and age-restricted videos can't be opened yet.
After the article is out
The page is put together from the text every time it's opened, so a change in the design reaches every article at once. When the author of the video signs in, the article becomes theirs to rewrite.