MediaDash

You know the content you like! But social media feeds distract you with other stuff. Instead of infini-scrolling, see all your entertainment in one place! MediaDash respects your time.

Comment 1

Inspiration

Most users of social media and streaming platforms utilize more than just one app. However, it's difficult for users to find the content they want to see because there's so much content dispersed across platforms, resulting in users wasting their time through doom scrolling across multiple platforms.

What it does

Taking inspiration from streaming apps like Spotify and Netflix, MediaDash is an innovative entertainment manager and recommender that brings all your media subscriptions and recommendations into one place, saving you time and enhancing your content discovery experience by utilizing AI to recommend content based on your watch history.

Features

  • Unified Dashboard: View content from multiple platforms in a single, user-friendly interface.
  • Personalized Recommendations: Get tailored content suggestions based on your preferences and mood.
  • Multi-Platform Support: Integrates with YouTube, Spotify, IMDB, and more.
  • Smart Search: Find content across all your subscriptions with a powerful search feature.
  • User Profiles: Customize your experience with genre preferences and platform selections.
  • Content Rating: Rate content out of 5 stars to improve future recommendations.

How we built it

We used:

  • React.js for frontend user interactions.
  • Langchain API to incorporate LLMs for agentic workflows.
  • Vercel for user-authentication.
  • Supabase for user data storage and communication between frontend and backend.
  • Google Gemini as the primary LLM model.

Tracks

  • best overall yay :3
  • the headstarter ai/ml prize
  • most likely meme startup
  • (WBUOT) would blow up on tiktok
  • (LAKOG) lowkey actually kind of good
  • (WPPOIL) would piss people off on linkedin
  • most cvrvey website
  • (IJAG) I’m Just A Girl
  • codedex i wanna learn stuff prize
  • wakaba prize
  • PearAI (YC24) prize
  • UI/UX most kawaii
  • build a tool that distracts you when ur trying to be productive
  • kanika mohan

Challenges we ran into

  • Providing recommendations that take into account previous watch history, mood keywords, and content metadata such as popularity, ratings, etc.
  • Incorporating multiple content sources and retrieving data for use in creating custom recommendations for each user.
  • How to enrich data being given by the user beyond basic mood keywords.
  • Seamlessly integrating APIs into agentic workflows without significant latency issues or AI hallucinations being produced.

Accomplishments that we're proud of

  • Utilizing AI to incorporate agentic workflows in the backend (agents for parsing user input, creating enriched search queries, parsing multiple recommendations from sources into different LLM responses and summarizing that into one cohesive response)

What we learned

  • The potential for LLMs to streamline content consumption.
  • The value in designing an app focused around being as user-centric as possible to encourage mass adoption.
  • The need for optimization when processing large amounts of raw data to ensure low-latency, high-quality AI responses.

What's next for MediaDash

  • Incorporating more platforms, content review sources, and social media for sourcing user data to further tailor recommendations.
  • Utilizing a custom LLM model for making recommendations
  • Gathering user feedback to improve UX, finding what sources are most preferred, and making the process as seamless as possible.

Built With

+ 5 more
Share this project:

Updates