科亿知识库 (KY KMS) is a document-centric knowledge management system built on Elasticsearch, designed for teams and organizations needing to organize, classify, retrieve, and share documents. It offers powerful full-text search, AI-driven insights, and robust access control, aiming to transform documents into valuable knowledge assets.

How It Works

KY KMS leverages a backend built with Spring Boot and Elasticsearch for its core search and indexing capabilities, supporting billions of documents and advanced retrieval features like keyword highlighting and in-result secondary searches. It integrates with LibreOffice and Tika for document processing and text extraction from various file types, including scanned documents. The system supports both local LLMs (via Ollama) and third-party AI models for features like summarization, classification, and natural language Q&A.

Quick Start & Requirements

• Installation: Docker deployment is strongly recommended. Pre-compiled installation packages are also available.
• Prerequisites: Java 8, MySQL 5.7+, Oracle 11g, or SQL Server 2017; Redis; Node.js 10.0+; Npm 5.6.0+ or Yarn 1.21.1+. Elasticsearch 7.6.1 and Libre Office 7.1.4 are required for backend functionality.
• Resources: Minimum configuration: 2-core CPU, 4GB RAM. Typical configuration: 4-core CPU, 8GB RAM.
• Documentation: Technical Documentation (Note: The provided link points to Jeecg-boot documentation, which KY KMS is based on).

Highlighted Details

• AI capabilities for document classification, summarization, and intelligent Q&A using local or third-party LLMs.
• Comprehensive full-text search across document content, attachments, titles, and keywords, with support for various file formats and online preview.
• Flexible access control and sharing mechanisms, including internal/external sharing, comments, and ratings.
• Local deployment option ensuring information security, with support for integration into platforms like DingTalk and WeChat Work.

Maintenance & Community

The project is actively developed, with the latest version (V2.0.0) released on October 25, 2024. A professional version is available for commercial use. Community interaction is facilitated via QQ group (782686853) and WeChat. The project is based on the Jeecg-boot framework.

Licensing & Compatibility

Licensed under the Apache License 2.0. Users are permitted to use and modify the software freely, but must retain copyright notices for derivative works.

Limitations & Caveats

The README mentions that the "open source version is suitable for personal learning and self-use, providing basic functions." Advanced features are reserved for commercial versions. The provided link for technical documentation appears to be for the underlying Jeecg-boot framework, not specifically for KY KMS.