The question
Working with private documents raises a practical question: can useful retrieval and inference stay on local infrastructure? Sentinel RAG explores that workflow.
The approach
A FastAPI backend and React/TypeScript interface connect document processing with local retrieval. Unstructured.io extracts content from PDFs and DOCX files; Ollama and FAISS support local inference and retrieval workflows. Metadata and structured processing make documents easier to search and analyze.
The result
A public Python project exploring local document intelligence and privacy-focused inference. Its architecture brings document preparation and local language models into one retrieval workflow. The repository documents the setup and current implementation.
Project information from the original portfolio and linked public sources. Explore the repository ↗