Podcastly

• 

  Home screen

• 

  Podcasts list

• 

  Episode

Inspiration

We’re all naturally curious, always picking up new things. But with the constant stream of media, it’s easy to feel overwhelmed. When I want to learn something, I try to fully immerse myself in it—it makes the process easier and more enjoyable. Podcasts seemed like a great way to do that, but when it came to research, they often skimmed the surface. They were engaging, but the depth just wasn’t there. That got me thinking: How can I make the most of my spare time while actually diving deep into a topic?

What it does

Podcastly. transforms written knowledge into personalized audio learning experiences. Users simply input their topic of interest, and our AI-powered platform generates an engaging podcast-style narrative. The current prototype focuses on web content, converting pages, articles into clear, structured audio narratives that adapt to the listener's preferred level of detail. Think of it as having a knowledgeable friend explain a topic to you, complete with proper context and engaging delivery.

How we built it

We developed our solution using a pipeline architecture that breaks down the complex task of content transformation into manageable steps. The system expands the user query to produce queries to find relevant information. Then it iteratively refines the information got from the web. We implemented this using Python, leveraging LLMs as primary engine and integrating with Elevenlabs services to create a seamless experience. We used techniques from Deep Research for navigating the web and collecting relevant information. The key was to let agent crawl the web and search for info on its own.

Challenges we ran into

The biggest challenge was striking the right balance between maintaining factual accuracy and creating engaging content. Integration of multiple sources of data via web search and tuning the system into producing

Accomplishments that we're proud of

Our biggest achievement was developing a prototype that effectively incorporates advanced deep research techniques, meeting high industry standards. This accomplishment highlights our team’s ability to blend cutting-edge technology with practical application. Additionally, we collaborated effectively to overcome complex challenges, resulting in an innovative and user-friendly solution.

What we learned

This project taught us valuable lessons about the intersection of content transformation and user experience. We discovered that creating engaging audio content involves much more than simple text-to-speech conversion - it requires understanding the rhythms of human speech, the importance of context, and the art of storytelling. We also gained practical experience in working with large language models and learned how to effectively combine different AI technologies to create a cohesive user experience.

What's next for Podcastly

Our vision extends beyond proof-of-concept. We plan to develop a more sophisticated research system that can synthesize information from multiple credible sources better, including academic papers and expert publications. We're also exploring features like: Voice customization for different types of content Interactive elements that allow listeners to dig deeper into specific topics Integration with learning management systems for educational institutions Development of domain-specific knowledge models for professional fields The prototype we built during this hackathon is just the beginning of our journey toward making knowledge more accessible and engaging for everyone.

Built With

• elevenlabs
• lovable
• openai
• python
• react