In-house Knowledge Search Application

Overview

A verifiable RAG-type web application that answers based solely on information existing in internal documents, with traceability from answers back to relevant source materials. Rather than a chat interface, it is designed around a verification workflow centered on question → search → evidence extraction → answer → citation → source confirmation.

Main Features

• Integrated AI Answers and Citations - Answers include document name, page, heading, and excerpt, with citations enabling navigation to relevant sections in the Document Viewer
• Access Control - Evaluates viewing department before search and excludes documents outside access scope
• Hybrid Search - Combines vector search and keyword search
• Document Version Management - Prioritizes newer versions within the same document family and excludes older versions from standard search
• Flexible Filtering - Enables search condition specification by category, department, document type, and date
• Evidence Visualization - Developer View displays search query, retrieved chunks, similarity scores, and answer generation stages
• Multi-format Support - Document registration for PDF, DOCX, TXT, Markdown, CSV (maximum 8MB)
• Follow-up Support - Supplements search queries with preceding conversation context and re-searches each time

Characteristics

• Returns "Cannot be confirmed from registered materials" when insufficient evidence exists
• Prevents hallucination by disallowing answers or speculation outside context
• Mitigates prompt injection by excluding imperative text from answer sources
• Manages API keys as server environment variables, not included on client side
• Demo mode does not transmit data to external AI