Problem Statement
Users need a feature that enables them to upload PDFs from their local devices, view them directly in the portal (either maximized or in a new tab), and ask questions about the content to receive AI-generated answers. The main challenges involve accurately extracting text from PDFs, ensuring fast and efficient storage and search of documents, and generating precise, context-aware answers without hallucinations.
From Documents to Insights
Our portal makes working with documents effortless. Users can upload PDFs, view them, and simply ask questions through a chat interface. Behind the scenes, files are securely stored, text is extracted with OCR (scanned or image-based PDFs can be read) if needed, and AI-powered embeddings are created using AWS Bedrock. Works well with large PDFs, so the user can just chat or prompt to get answers instead of scrolling through hundreds of pages. With OpenSearch ensuring lightning-fast retrieval, questions are instantly answered with clear, context-based responses.
In short — secure storage, powerful AI, and a simple interface that helps you find answers in seconds.
Flow Diagram
Solution
- Upload & View: When a user uploads a PDF, the system stores it securely in S3. The document can be viewed directly on the portal or opened in a new browser tab with a PDF viewer.
- Text Extraction & Indexing: The system extracts text (using parsers or OCR for scanned files), splits it into chunks, and generates embeddings with AWS Bedrock. These chunks, along with metadata, are stored in OpenSearch for quick semantic search.
- Question Answering: When the user asks a question, the system generates an embedding for the query, searches OpenSearch for relevant chunks, and passes them to a Bedrock LLM. The model generates an answer grounded in the document content.
- Answer Presentation: The answer is displayed to the user along with references (page numbers or text snippets), allowing easy navigation to the exact part of the document.
The Chat with PDF feature improves productivity by enabling users to ask direct questions and get quick, context-based answers without reading entire documents. It enhances accuracy by grounding responses in the actual text, ensures smooth user experience through an interactive portal, and leverages AWS services for scalability and security. Flexible enough to handle different document types and formats, this solution is built to adapt and grow.
In conclusion, all data is handled within our AWS environment and users don’t need to worry about data security. AWS does not provide any out-of-the-box APIs for Chat with PDFs, making our solution a secure, scalable, and user-friendly way for users to engage with their documents more intelligently and efficiently.