So you have forty videos and a blog nobody has updated since March.
Everyone keeps telling you to repurpose. One video becomes a blog post, five social posts, a newsletter, and apparently a podcast episode too. The maths always sounds great in the thread where you read it.
Then you actually sit down to do it and it takes four hours, the blog post reads like a robot transcribed a man talking, and you quietly never do it again.
The advice isn't wrong. The execution advice is just missing.
The mistake everybody makes on the first attempt
You paste the transcript into ChatGPT and ask it for a blog post.
What comes back is technically an article. It has headings. It covers the same ground the video covered. And it is completely dead, because a transcript of someone talking is not an outline of an argument, and asking a model to convert one into the other without telling it which parts matter gets you a smooth average of everything you said.
Spoken content rambles. That's fine, that's what makes it watchable. You circle a point, drop it, come back to it eleven minutes later with a better version. A viewer follows that happily. A reader will not.
So the transcript is raw material. It is not a draft.
Once that clicks, the whole workflow changes, and it gets faster rather than slower.
Step one, get the text out
This part should take two minutes and no thinking.
Most transcript tools now handle a full video without an account. Using Vomo as the example, because it doesn't put the basic version behind a signup wall:
Copy the video URL. Address bar is fine. Unlisted videos work, which matters if you're processing something before it goes public.
Paste it and click Generate Transcript, It's Free. No account needed at this point.
Check the video it pulled back. It shows the title and duration so you can confirm it grabbed the right one. Ten seconds. Do it anyway, because copying the wrong link off a channel page happens more than anyone admits.
Hit Transcribe. Processing runs. A ten-minute video is quick, a two-hour livestream takes a few minutes, and there's no cap on the length of a single video, which is the thing that usually breaks free tools.
Read what comes back before you export. You get the transcript plus a summary with timestamped chapters.
Sign in when you want the file. Free account, then TXT, DOCX, PDF or SRT.
That last step is where most people stop reading and start pasting into an AI. Don't. The next step is where the whole thing actually works.
What the transcript is actually for
It's for finding out what you said.
Not what you planned to say. What you said. Those are different, and the gap between them is where the good material lives.
Read the summary and the chapter list first. Two minutes. What you're looking for is the moment you went off script, because that is almost always the best part of the video and it's almost never in your outline. Somebody asked a question and you answered it better than you've ever answered it before. You made an analogy you'd never used. You contradicted something you say all the time.
Mark those. Three or four per video, usually.
Now you have something to write from. Not "the video, in text," which is a shapeless forty minutes. Four specific claims, in your own words, with the exact phrasing you used when you weren't thinking about phrasing.
That's the difference between a repurposed post that reads like a transcript and one that reads like you.
Five outputs, ranked by whether they're worth it
Everyone lists the same five. Nobody tells you three of them aren't worth your Tuesday.
The blog post. Worth it. But write it from the four marked moments, not from the transcript. Structure it as an argument rather than a chronology. It will share maybe 60% of its content with the video and that's correct, because the 40% you cut was throat-clearing and the parts where you looked at your notes.
The newsletter. Worth it, and it's the easiest one. One marked moment, expanded, with a link to the video at the bottom. Twenty minutes. This is the highest ratio of output to effort in the entire list and it's the one people skip because it feels too small to count.
Social posts. Worth it if you stop trying to write five. Take the single best line, the one that surprised you when you read it back. Post that. Five mediocre posts from one video performs worse than one good one, and it costs you four times as much attention.
The podcast version. Usually not worth it. Stripping the audio off a video where you're pointing at things on screen produces a confusing audio file. If the video was talking-head, fine, do it. If it had any visual component, you're publishing something broken.
The "SEO article" version. Depends entirely on whether you have anything to add. A post that just restates the video will sit at position 40 forever. A post that answers the questions the video raised but never addressed has a chance. The transcript tells you which questions those are, because you can see where you said "I'll come back to that" and then didn't.
The back catalogue is the real money
Everything above assumes you're processing this week's video, which is the obvious move and not the valuable one.
The valuable one is the forty videos already sitting there. Most creators have a back catalogue that stopped earning the day the algorithm moved on, and the reason nobody mines it is that rewatching your own material from two years ago is genuinely unpleasant work that produces nothing for hours at a stretch. Reading it is a different proposition entirely, because you can process a forty-minute video in four minutes of skimming and know immediately whether there's anything in it worth rebuilding.
Run ten old videos. Read ten summaries. You'll find two that were ahead of their time and are now exactly what people are searching for, one that you'd argue against today which is its own excellent post, and seven that deserve to stay buried.
That's a morning. It's also, realistically, a quarter's worth of blog posts sitting in work you already paid for.
Interviews are a different job
If your video has two people in it, the workflow shifts.
Speakers get labelled automatically and you can rename them afterwards, which sounds cosmetic and isn't. A named transcript is quotable. You can pull a line, attribute it correctly, and drop it into a post without going back to the video to check who said what.
There's also a chat function that lets you ask the transcript questions directly. On a ninety-minute interview this earns its place immediately. "What did she say about pricing" beats scrolling for six minutes. On a fifteen-minute solo video it's overkill and you should just read the thing.
Worth flagging: multi-speaker audio comes back messier than one person to a camera. Two people talking over each other will need a manual pass. Budget for it on panels.
Where this breaks
A YouTube transcript generator is very good at one thing and it will not save you from the rest.
Accuracy sits around 95% on clean audio across fifty-odd languages, which is high enough that you'll stop checking. Then you'll publish a quote with a misheard proper noun in it. Names, brands, technical terms and numbers are where it slips. Check anything you're putting in quotation marks.
Heavy background noise degrades it. A conference recording with room tone and a bad lapel mic is a hard case, and the transcript will read as slightly-off English rather than as obviously broken, which is worse because you won't notice.
And the free tier is 30 minutes of transcription a week. Fine if you publish weekly. Not fine if you're working through a back catalogue in one sitting, which is exactly what you'll want to do the first week you try this. Unlimited runs $1.92 a week. Cheap enough that the decision isn't interesting.
A realistic week
Here's what this actually looks like when it's working, and it's smaller than the threads make it sound.
Monday, you publish the video. Same day, run it through and read the summary. Fifteen minutes total, and you mark your three or four moments while the video is still fresh.
Tuesday, write the newsletter from one of them. Twenty minutes.
Wednesday, the blog post. An hour, maybe ninety minutes, and it's real writing rather than transcript-wrangling because you already know what the four points are.
Thursday, one social post from the best line.
That's it. That's the whole system. Under three hours a week, and the output is a post, a newsletter and a social item that don't read like each other or like the video.
Compare that to four hours of fighting a transcript into a shape and abandoning it.
The bit that matters
Repurposing fails when you treat the video as the finished thing and everything else as a copy of it.
It works when you treat the video as the first draft you happened to say out loud. The transcript just makes that draft readable enough to edit.
You already did the hard part. You just left it in a format nobody can work with.