Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{197493,
        author = {ANTONY PRAVEEN E IT},
        title = {FAST AND MODULAR RAG FRAMEWORK FOR LONG PDF DOCUMENTS},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {7018-7023},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197493},
        abstract = {Retrieval-Augmented Generation (RAG) systems have significantly improved knowledge-based NLP tasks, but existing frameworks like LangChain and Haystack face challenges when handling long, structured PDF documents. Traditional approaches rely on simple text chunking, which ignores document layout, hierarchy, and tables, leading to loss of context and lower retrieval accuracy. To address this, the proposed research introduces a fast and modular RAG framework designed specifically for such documents, starting with structure-aware PDF parsing that preserves text, layout, and tabular data. It uses intelligent, layout-based chunking to maintain logical flow, along with lightweight sentence embeddings generated by all-MiniLM-L6-v2 and efficient semantic search through FAISS indexing. A key feature is the integration of the Model Context Protocol (MCP), which connects the retrieval system, database storage, and locally deployed models via Ollama, ensuring that both text and table data are combined effectively before generating responses. As a result, the system produces more accurate, context-aware outputs with lower latency and resource usage, and experimental results show that it offers better retrieval performance and scalability, making it well-suited for enterprise-level document processing.},
        keywords = {},
        month = {April},
        }