Inspiration

Healthcast transforms a simple natural language prompt (e.g., “I’m 70kg, I jog twice a week, and I want to eat healthier”) into: Personalized fitness & nutrition plans A motivational podcast episode with actionable tips A concise weekly markdown guide

Instead of reading dry instructions, users can listen, learn, and follow along—making the journey engaging and sustainable.

What it does

Natural Language Input → Users type or say their lifestyle and goals. Smart Information Extraction → Regex, spaCy, and NLP models process the input. Personalized Predictions → Random Forest for tailored workout plans XGBoost for customized meal suggestions Creative Content Generation → Predictions are transformed into scripts and podcasts with Gemini + Murf. Engaging Outputs → Podcast file (podcast.mp3) Weekly plan (weekly_plan.md) Seamless Web App → Built with Streamlit for an intuitive experience.

How we built it

Challenges we ran into

Designing smooth NLP pipelines that balance structure with flexibility. Ensuring ML models generate practical, realistic recommendations. Making the podcast output engaging, motivational, and natural. Integrating multiple components (NLP, ML, TTS, UI) into one streamlined workflow.

Accomplishments that we're proud of

Built an end-to-end AI-powered app that connects NLP, ML, and TTS seamlessly. Created personalized workout and meal plans that actually make sense for users. Generated podcast-style audio that feels motivating and human-like. Designed a smooth Streamlit app for an intuitive user experience. Combined technical innovation with creative storytelling to make health guidance engaging.

What we learned

How to integrate NLP, ML, and generative AI into a unified workflow. The importance of balancing technical accuracy with user engagement. That health advice is more effective when delivered in a narrative, human-like way. Collaboration across AI, software engineering, and design disciplines can produce something more impactful than the sum of its parts. Building for sustainability and motivation is just as important as building for accuracy.

What's next for Healthcast

Expanding personalization with more diverse fitness & nutrition datasets. Adding voice input for an even more natural experience. Supporting multilingual podcast generation. Building long-term tracking features to monitor progress and adjust plans. Exploring integrations with wearables (smartwatches, fitness trackers).

Built With

  • custom-tts
  • fitness-datasets
  • gemini
  • murf
  • python
  • random-forest
  • regex
  • spacy
  • streamlit
  • xgboost
Share this project:

Updates