In2ne - AI-Driven Music Recommendation

A dual-mode app using AWS Rekognition, OpenAI, Pinecone vector search, and a multi-agent system to recommend on-trend, contextually relevant music based on a user-uploaded image.

In2ne analyzes an uploaded picture using AWS Rekognition to label elements. OpenAI transcribes these into a creative story, which is then analyzed by an AgencySwarm multi-agent system to assign matching audio features. It uses Pinecone for vector search and a k-cluster ML model (SageMaker-deployed) for similar song recommendations. For content creators, it includes a trend analyzer agent that scrapes viral music data from platforms like TikTok and YouTube via APIs. The frontend is Next.js/Tailwind/TypeScript, with a Python/Flask backend and extensive AWS integration (Rekognition, S3, SageMaker, Polly).