Inspiration

We all have busy lives and most of us love to read books and listen to long podcast. But how could we possibly find enough time for all these exciting books, audiobooks and interviews?

What it does

Compressa.NET makes it possible, or the very least it makes it a lot more efficient. What if I told you that you have a companion who reads the books on your backlog for you, and listens to your favorite podcasts and 4-hour-long Lex interviews?

Now you have indeed! Compressa summarizes the audiobooks, podcasts* and YouTube* channels for you! (*coming soon :))

How we built it

I used the latest tech I knew about: .NET 7 (released in this November), .NET MAUI (released in this October), AssemblyAI's summarizer (release in this October). They were probably a risky choice, and I run into some surprising problems, but all in all, it turned out to be okay.

The backend runs with ASP.NET, the frontend with .NET MAUI. I wrote up a detailed technical discussion on my GitHub (see: https://github.com/compressa-net/compressa#technical-details)

Challenges we ran into

There is one important aspect of .NET MAUI I didn't know about beforehand: although it is platform-independent, it can not be published as a web application, it must be an app installed on the system.

Its platform independency isn't mature either: as I tested the releases they either need a signing key (on Windows and Mac), or they crash (on Android) for some reason. Probably just because it's an early tech.

The other challenge I had is that not all the AI APIs have C# libraries, certainly not in the NuGet library. That caused me delay, because I had to write them myself.

Publishing to the Microsoft Store is really not that straightforward as I thought, it was an emotional roller-coaster.

Also, I nearly forgot how difficult to go to sleep when you are excited and your brain can't turn itself off! :)

Accomplishments that we're proud of

In the end I have a great little app, I think. I thought that it would be impossible to do it during the weekend without a solid frontend developer, but it turned out ok. GitHub Copilot helped A LOT!

What we learned

I learned that C# could be more popular among developers to have more support, and that early tech can always surprise you with funny bugs and other quirks.

What's next for Compressa.NET

Countless ideas formulated in my head over the weekend, and I plan to implement them with the help of others, if I win any of the prizes.

Just a few:

  • Integrate translations
  • Integrate Uberduck for voice generation (even by cloning the original audiobook voice)
  • Have a subscription model and share the revenue with the authors based on their popularity
  • Create a chatbot that can chat with you about the book and many many more...

Built With

Share this project:

Updates