PodQuirk | Devpost

Title
what is bitcamp
entertainment

Inspiration

I wanted to experiment with ElevenLabs to see how far I could push realistic AI-generated voices. It seemed like a fun challenge to create something functional and creative, and podcasts felt like a natural fit.

What it does

PodQuirk lets users pick a topic or category, and in under a minute, it generates a podcast episode that talks about that subject. It combines real-time news and smart summarization with AI narration, giving users a quick, engaging audio experience.

How I built it

The frontend is built with Next.js, the backend with Flask. I use OpenAI to generate the podcast script, ElevenLabs for lifelike TTS audio, and the Newspaper3k API to pull relevant news content. AWS handles caching and storage to reduce unnecessary API calls and speed up repeat requests.

Challenges I ran into

Managing API usage was a big one! I didn’t want to burn through credits every time someone clicked “generate.” So I implemented a caching mechanism using AWS to store previously generated episodes. Another challenge was fine-tuning the prompt for script generation so that it would sound natural and match the tone of a real podcast. Setting up and managing ElevenLabs also had a bit of a learning curve.

Accomplishments that I'm proud of

I was able to generate a full podcast in under a minute, from topic selection to audio output. That includes gathering relevant news, scripting, and producing it all with high-quality narration.