Hello,

I am a developer building a strictly non-commercial movie recommendation system for my personal portfolio.
The application will be deployed publicly, but it will have no ads, no monetization, and no commercial intent whatsoever.

To implement a hybrid search features, I plan to retrieve movie metadata (titles, IDs, actors, watch providers, and overviews) via the TMDB API.

I would like to clarify a couple of points to ensure full compliance with your Terms of Use:

1. Data Storage for Hybrid Search (Relational + Vector DB):
   I understand that harvesting data for permanent mass storage is restricted, but temporary caching for performance is permitted.
   app requires storing the retrieved metadata in a relational database for basic display, and in a vector database to enable semantic search.

This setup is solely for application performance and functionality, not for creating a separate data clone. Is this database architecture compliant with your caching policy?

2.AI/ML Clause & Text Embeddings
Your terms restrict using TMDB content to train machine learning or AI models.
My application does NOT train, fine-tune, or optimize any AI models. It simply converts the movie overviews into text embeddings to power a retrieval-augmented generation (RAG) / semantic search system.
Does this use case violate your AI/ML clause?

I will strictly follow all TMDB Branding and Attribution Requirements, including prominently displaying the official TMDB logo and the mandatory disclaimer.

Thank you for your time and guidance.