Life Update: I Just Sold My App (Part 1)

Hey,

Figured I hadn't posted in a while, so just a quick update.
I’m happy to announce I’ve sold my app after only 60 days of building it.

If you remember, I was previously working on an AI sidebar for YouTube that allowed you to “chat” with the video.
It was named ChatWithYT.

The initial product looked a lot like Eightify, and while I was doing some research on how I should position my app (because I didn’t want to just be a mere clone), I stumbled upon the StudyTok niche where there were a few noticeable outliers:

  • StudyFetch
  • YouLearn.ai
  • TurboLearn.ai

Each of those platforms offered a very similar feature, but they also had strong similarities with existing tools like NotebookLM, RAG-based tools, and were way more appealing for students with simple Whisper features like “record your lecture” in order to give accurate transcriptions of any live class.

In addition, they also had integrated quiz and flashcard generation modules and, in the case of YouLearn.ai, ChatGPT Advanced Voice Mode-like agents you can interact with.

All of this made me want to pivot my tool to an AI study tool, because I wasn’t really sure what to make of my extension.

My Problem

I didn't have a well-defined audience.
Was it for students?
Working professionals?
Researchers?

  • YouLearn.ai, StudyFetch, TurboLearn.ai → 100% focused on students.
  • Anara, NotebookLM, Eightify → aimed more at researchers or professionals.

Where did I fit??

Building the App

I was already pretty proud of my little tool: anytime I was watching a YouTube video that felt a little boring, and I wanted to squeeze the juice out of the content quickly, I used my tool and didn’t need to watch the whole video. So I personally found it pretty useful.

Looking back, I fell a bit into the overthinking trap, because I could have marketed my product as-is:
“Stop wasting time sitting through boring YouTube lectures.”
I could have shipped my extension as it was.

Instead, I fell into the feature trap and spent about a month rebuilding my entire app as a Next.js app with a FastAPI backend, and the following functionalities:

  • AI chat with model switching (DeepSeek R1, Claude Sonnet 4, Gemini 2.5 Pro, Gemini 2.5 Flash via OpenRouter)
  • Note-taking app using TipTap and integrated generative AI to create drafts faster
  • File imports (DOC, PDF, PPT, Markdown, TXT)
  • YouTube video imports (reused some logic from the ChatWithYT extension)
  • Voice recording and live transcription of lectures via a Whisper model
  • Summaries, study sheets, quizzes, and flashcards generation

The product was somewhere between Anara and YouLearn.ai.

How Did I Market My Product?

While building, I created multiple TikTok accounts. I managed to get my hands on a total of 6 phones (mostly iPhones), paid for a few VPNs (I used HMA and Mullvad), and together with my girlfriend we started creating TikTok content.

There was nothing innovative — we just identified and copied winning formats in the U.S. and localized them for a Vietnamese audience.
(The reason for going Vietnamese, I’ll explain later.)

After several days of posting, we finally got our first users and quickly received a few first payments.

The whole process of creating TikTok content was just:

  1. Grab a ton of pics from Pinterest
  2. Use ChatGPT/Claude to generate the captions
  3. Use Canva to do the editing

As easy as that.

One good trick is to have multiple hooks, middle parts, and CTAs to mix & mash.
It’s super easy, once you identify a winning format, to just recycle it over and over again.

We eventually hit 200 users in less than 2 weeks.

From Building to Selling… in 60 Days?!

As mentioned a bit above, I was mainly advertising to a Vietnamese audience.
The reason is that I had previously met a potentially interested buyer before I even started building. So, I sort of pitched him the project, and he was very interested in acquiring it.

We met a couple of times in Ho Chi Minh, had a few phone calls, and we agreed on a number quickly.

So, after a month of building—once I was done with the app’s core functionalities—we met again and agreed to do the funding.

(I’ll continue this story in another post.)

julien jung#now