MyRSSPress

• 

  1 What are you interested in?

• 

  2 AI Feed sugestion

• 

  3 Here is the newspaper example

• 

  4 You can save your own newspaper and also see other people's newspapers

Inspiration

I wanted to resurrect two forgotten technologies: RSS feeds and newspapers. In an era dominated by social media algorithms, I missed the open, decentralized nature of RSS and the curated reading experience of traditional newspapers. MyRSSPress combines both with modern AI to bring back user-controlled, personalized information consumption.

What it does

MyRSSPress generates AI-powered personalized newspapers from RSS feeds. Users enter their interests, and Claude AI suggests relevant feeds. The app fetches articles, ranks them by importance, and displays them in a beautiful newspaper-style layout. Users can save and share their newspapers, creating a community of curated content.

How we built it

I built MyRSSPress entirely with Kiro AI as my development partner. Frontend uses Next.js 15 with Tailwind CSS for the newspaper layout. Backend runs on AWS Lambda with Hono framework. Claude 3 Haiku powers feed suggestions and article importance ranking. DynamoDB stores newspapers, and Amplify handles deployment. Everything is serverless and scalable.

Challenges we ran into

The biggest challenge was validating AI-suggested feed URLs—Claude sometimes suggested non-existent feeds. I solved this by implementing HEAD request validation before showing suggestions. Another challenge was designing the newspaper layout algorithm to handle varying article counts and image availability while maintaining visual balance. I also had to handle legacy newspapers without stored articles by implementing on-demand regeneration.

Accomplishments that we're proud of

I'm proud of the seamless UX flow with loading and success animations that guide users naturally. The newspaper layout algorithm creates visually appealing layouts regardless of content. The entire app was built in collaboration with Kiro AI, demonstrating effective human-AI pair programming. I also achieved full i18n support (Japanese/English) and implemented comprehensive security checks with Gitleaks.

What we learned

I learned that AI pair programming with Kiro dramatically accelerates development—features that would take days were completed in hours. I discovered the importance of validating AI outputs (like feed URLs) before presenting them to users. I also learned that modern serverless architecture (Lambda + DynamoDB) can handle complex applications while keeping costs low and scalability high.

What's next for MyRSSPress

Next, I plan to add user authentication for personalized newspaper collections, implement email delivery for daily digests, and add social features like following other users' newspapers. I also want to expand AI capabilities to summarize articles and detect trending topics across feeds. Finally, I'll add export to PDF for offline reading and printing.

Built With

• amazon-web-services
• amplify
• kiro
• lambda
• next.js
• typescript