Allan Nascimento

Hey everyone, this is Allan, and in today's video I already got an improvement. Last video my background was too dark, so today I am using my cell phone to record my camera instead of my notebook's webcam.

Today I want to talk about something that I did the day before yesterday, on Saturday. I will talk about how I extracted content from the videos that I have already recorded, and how I plan to distribute it in a way that is content created by me, but that I can schedule and publish in volume.

The problem I had

First I want to talk a little bit about a problem that I had. I planned to make daily posts, and I planned to make them based on the morning pages that I write in my handbook. But it was not working, because my morning page is usually something really personal. I write a lot of planning and ideas that I have, and I realized that it was not good content to try to grow an audience.

I also started publishing the same posts on X and on Threads. I realized that the things that I planned to do, the things that I wrote here, were getting so few views that no one was actually seeing them. And when I published stories of what I did in the past, of what happened to me, they got a lot more views.

I have no audience for now, but the algorithm would show my content a little bit more when it was things that happened to me instead of things that I plan to do in the future.

The videos as a source

So instead of getting the ideas from the handbook that I write daily, I decided to use as a source the videos that I already recorded in this channel. I have more than 100 videos. They don't have a lot of views, and I know that I need to improve a lot in all those videos. But they are a better source than the content that I tried to create before.

How I extracted it

So, how did I extract it? I basically got all the transcriptions from the 103 videos, because I had already saved all of them.

I decided to take only literal quotes, because I don't like the idea of AI writing things for me. I prefer to record a video, to struggle a little bit talking and juggling my ideas. It took me a second to remember the verb juggling. I don't like to use AI to create text that I will publish. So I prefer to record the video and get literal quotes from the videos, because it is something that I said.

I also found a lot of errors in the transcripts, but all of them were fixed. And I was only able to do that at scale because I am using AI. I used Claude Code with some agents to do that work, and I transcribed the videos with Groq, using the Whisper Turbo model.

What I got

The result is that I got more than 600 literal items of content from my videos. 500 of them are short posts, short phrases that I will publish with time. And 67 were threads, posts that are longer. I also take these threads and publish them as carousels on my Instagram that is in English and also on my Instagram that is in Portuguese. So I have two accounts right now.

I also got some cuts from these videos. Actually, I got cuts only from the last video, because I want to get better cuts, I would say. So after I record this video, I will get cuts from it, but not from the past 100 videos.

Scheduled until March

I already have posts scheduled until March of next year. I will be 28 years old in March, and I will still be publishing posts about the ideas that I had this year: texts and short paragraphs about the things that I recorded this year.

But I will not only plan things that are from the past 100 videos. I set the amount of 10 posts a day for X and Threads. I will publish more or less 4 new pieces of content every day based on the videos that I record, and 6 of them will be the ones that I already extracted from my past videos. These are all my ideas. They were things that I thought about and things that I talked about in this channel.

The carousels on Instagram will be published until next month. I plan to publish them in 30 days, because I think that carousels are better for growing an audience on Instagram. So it will be a test for me. All of them are already queued on a server that I have, and they are all scheduled already, so it will be an automatic process from now on. I only had to struggle a little bit with recording the ideas, but now the distribution of that content is already automatic.

One video as input

And that's the plan from here. In short, I planned that every new YouTube video is the input and all the rest is derived from this input. So I have one video that will be the input, and a lot of content from it that will be the output.

I found that it is a lot easier for me to record a video than to try to write posts every day. So I think that I found a really good way to create content.