PodSnap.AI

• 

  The hero section of the website.

• 

  The dashboard showing currently followed podcasts.

• 

  An explanation of the main steps when using PodSnap.AI.

• 

  The dashboard showing the UI for finding and following podcasts.

• 

  Some of the featured podcasts that can be followed for free.

Inspiration

I was catching up with a former colleague who said he loves podcasts, that there are too many to keep up with, and that they've searched for a solution but couldn't find one.

What it does

PodSnap.AI app keeps track of the podcasts you follow, detecting when new episodes come out. It then automatically generates text and audio summaries using the latest AI technology and sends you an email notification.

This lets you stay up-to-date with your favorite podcasts and easily decide which episodes to listen to.

How I built it

I built the app using the NextJs-TypeScript-MUI-Postgres stack. The backend also talks to various AI services to generate text and audio summaries, as well as AWS S3 to store audio files.

Challenges I ran into

NextJs is not great at performing async tasks, which this service needs to do to achieve a good UX. Since task queue systems are overkill for this, I call a GCP Function, which does a callback to my NextJs: this way, the user gets immediate feedback while the task can be done async using the same codebase.

Accomplishments that I'm proud of

It's difficult to keep the costs low enough for such a service to make sense. I am proud that, after a lot of experimentation, I was able to achieve this, while at the same time supporting 4.2M+ podcasts!

What's next for PodSnap.AI

It's still early days for PodSnap.AI. There are still many improvements that I would like to make, but the focus at the moment is to gather feedback from the next wave of users and iterate accordingly.

Built With

• ai
• gcp
• nextjs
• postgresql
• postmark
• vercel