Project Spear: An AI-Powered Digital Memory Assistant Using Retrieval-Augmented Generation with Secure Access Control
Information kept in personal and organisational documents is scattered across devices and platforms, and conventional keyword search and general-purpose chatbots cannot answer questions about it reliably. This paper presents Project Spear, an AI-powered digital memory assistant that answers natural language questions from a user’s own documents using Retrieval-Augmented Generation (RAG). The proposed framework works in two phases. In offline indexing, uploaded documents are loaded, preprocessed, split into chunks, converted into embeddings and stored in a vector database (FAISS or ChromaDB). In the online query phase, the question is embedded, the most similar chunks are retrieved and optionally reranked, combined with the query into a context, and passed to a large language model that generates a grounded answer with optional citations. Authentication and access control, a React.js frontend, a FastAPI/Flask backend and a PostgreSQL/SQLite store for users and metadata complete the system. A survey of twenty recent RAG studies on evaluation, long-context handling, retrieval reliability, and security and privacy shows that none offers a complete, secure assistant for personal digital memory, which motivates the design. This paper reports the design phase of the project (architecture, data flow and requirements); experimental results are not yet reported.
Authors
- Muhammed Shammas Muhammed Suhail TS
- Muhammed Shahid KM
- Niveditha P S Thanseeh PP
Publication Details
- Journal
- International Journal of Advanced Research in Science Communication and Technology
- Published
- 2026-11-21
- DOI
- https://doi.org/10.48175/ijarsct-40915
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 0.00