Marketing Assistant

This project, the AI Marketing Assistant, solves the problem of generic AI content by implementing a robust Retrieval-Augmented Generation (RAG) pipeline. Unlike standard chatbots, this system doesn't rely solely on the LLM's pre-trained knowledge. Instead, it allows users to upload their private, proprietary documents—such as brand style guides, detailed product specifications, and internal case studies.These documents are processed in the backend, chunked, and transformed into vector embeddings using a powerful embedding model. The resulting vectors are securely stored in Qdrant, a high-performance vector database. When a user requests content (e.g., "Write a tweet about our new product"), the query is used to perform a similarity search against the Qdrant knowledge base. The most relevant document chunks are dynamically retrieved and "stuffed" into the prompt before being sent to the Gemini LLM. This process grounds the AI's response in the user's truth, ensuring the generated marketing copy is not only creative and engaging but also factually accurate and perfectly aligned with the brand's unique voice and current product details. The system uses FastAPI and PostgreSQL for a secure, scalable backend architecture.