1. Project Overview

Keyi Knowledge Management System (KYKMS)

An intelligent Q&A knowledge base that supports integration with large language models, including Deepseek. Features full-text search powered by Elasticsearch and comprehensive file management. Provides powerful and flexible organizational permission management, precise full-text and multi-dimensional search, online file preview, version control and rollback, mobile support, DingTalk/WeCom integration, rich extension APIs, third-party integrations, knowledge distribution, and multiple knowledge-sharing and communication methods.

2. Open Source License

Licensed under Apache-2.0

3. Screenshots

System Preview (Intelligent Edition)

Public Search

Intelligent Q&A Search HomepageSearch ResultsAdvanced SearchKnowledge TopicsKnowledge Details

Knowledge Sharing 

Admin Console

Data Statistics Knowledge Management 

Authorization Management

Editing and Creation 

4. Feature Overview

Project Introduction:

• If you are looking for a knowledge base system to organize and categorize documents, share them securely within your team to make full use of these resources, or share documents externally, Keyi Knowledge Management System is the ideal choice. Powered by AI, it becomes your second brain.
• Keyi Knowledge Management System can be used wherever files need to be managed, categorized, collected, searched, shared, and discussed, as well as in scenarios requiring big data management and retrieval. It is particularly suitable for knowledge-intensive organizations and organizations with extensive historical document archives.
• Edition Information: The open-source edition is suitable for personal learning and individual use. It provides essential features, has no hidden restrictions, and runs as a complete system. The commercial edition offers more powerful features and technical support. Please contact our sales team for details. The standalone edition is portable software that works immediately after downloading, with features comparable to the Professional Edition.
• Our Development Strategy: The open-source edition makes the product accessible to more customers and users. Revenue from advanced editions and technical services helps sustain the open-source edition.
• Keyi Knowledge Management System helps you quickly transform files into knowledge. Our mission is: Organize files simply, create value through knowledge.

Why Choose Keyi Knowledge Management System?

• With over three years of product operation and validation by hundreds of customers, our system is backed by a professional technical team, giving you peace of mind throughout its use.
• Powered by large language models and integrated with your local knowledge base, it creates an intelligent knowledge assistant that understands your needs. Let AI categorize, tag, and summarize your knowledge, and retrieve precise, concise information from your local knowledge base through intelligent Q&A.
• Supports locally deployed large language models (such as ollama), as well as integration with mainstream third-party models, including ChatGPT, Tongyi Qianwen, Zhipu AI, and ERNIE Bot.
• Flexible knowledge permission management with extensive customization options. Also supports flexible sharing and communication modes, including internal and external sharing, ratings, and comments.
• Knowledge review supports workflow management (based on Flowable) and customizable processes.
• Built on the powerful Elasticsearch search engine, it delivers advanced search capabilities and comprehensive search features. It supports flexible cluster expansion and can handle up to tens of billions of records.
• Comprehensive content search covering attachment contents, titles, keywords, and knowledge summaries. Supports secondary searches within results, advanced combined searches, exact matching, and keyword highlighting.
• Comprehensive file format support. Full-text search supports Office documents, text files, PDFs, and mind maps. Online preview supports Office documents, PDF, OFD, text files, mind maps, images, audio, video, and more.
• Automatically extracts text from images within attachments, making it easy to extract knowledge from scanned documents. Extracted content also supports full-text search and intelligent Q&A.
• Online previews support customizable watermarks to prevent screenshots of sensitive information from being leaked.
• Optimized for mobile H5 and supports integration with DingTalk and WeCom, including single sign-on and knowledge distribution through third-party systems.
• All capabilities and services are deployed locally, requiring no internet connection or external services, meeting enterprise information security requirements.
• Excellent performance. Minimum configuration: 2 CPU cores and 4 GB RAM. Recommended configuration: 4 CPU cores and 8 GB RAM.

5. Tech Stack

KMS Feature Mind Map

KMS Technical Architecture Diagram

Technical Architecture:

Technical Highlights

• Mainstream technologies, a mature ecosystem, and easy maintenance
• Frontend and backend separation, clear API documentation, and easy secondary development
• The backend uses the Elasticsearch search engine, allowing distributed deployment according to requirements

Development environment

• Language: Java 8
• IDE (Java): Install the Lombok plugin in IDEA / Eclipse
• IDE (Frontend): WebStorm or IDEA
• Dependency Management: Maven
• Database: MySQL5.7+ & Oracle 11g & Sqlserver2017
• Cache: Redis
• Node Js: 10.0 +
• Npm: 5.6.0+
• Yarn: 1.21.1+

Backend

• Base Framework: Spring Boot 2.3.5.RELEASE
• Persistence Framework: MyBatis-Plus 3.4.1
• Search Engine: Elasticsearch 7.6.1
• Document Processing and Conversion: LibreOffice 7.1.4
• Text Extraction: Tika 1.17
• Security Framework: Apache Shiro 1.7.0, JWT 3.11.0
• Database Connection Pool: Alibaba Druid 1.1.22
• Cache Framework: redis
• Logging: logback
• Others: fastjson, poi, Swagger-ui, quartz, lombok (to simplify code), etc.

Frontend

• Vue 2.6.10,Vuex,Vue Router
• Axios
• ant-design-vue
• webpack,yarn
• vue-cropper - Avatar Cropping Component
• @antv/g2 - Alipay AntV Data Visualization Charts
• Viser-vue - Implemented using the antv/g2 wrapper
• eslint，@vue/cli 3.2.1

Functional Modules

├─ Search Homepage
│  ├─ Basic Search
│  ├─ Advanced Search
│  ├─ Exact Search
│  ├─ Secondary Search Within Results
│  ├─ File Preview
│  ├─ Knowledge Version Control
│  ├─ Knowledge Details (including knowledge sharing, distribution, ratings, and comments)
│  ├─ Browse by Directory and Category
│  ├─ Personal Workspace
├─ Personal Center
│  ├─ My Knowledge
│  ├─ My Favorites
│  ├─ My Shares
│  ├─ My Messages
│  ├─ Change Password
├─ Knowledge Management
│  ├─ Create Knowledge
│  ├─ Edit Knowledge
│  ├─ Files Pending Review
│  ├─ Reviewed Files
│  ├─ Workflow Tasks
│  ├─ File Statistics
├─ Knowledge Base Configuration
│  ├─ System Parameters
│  ├─ Category and Tag Definitions
│  ├─ Security Classification Definitions
│  ├─ Directory and Category Definitions
│  ├─ File Data Masking Rules
│  ├─ Carousel Management
├─ Knowledge Permission Management
│  ├─ Additional Authorization Settings
│  ├─ Security Classification Access Settings
├─ System Management
│  ├─ User Management
│  ├─ Role Management
│  ├─ Menu Management
│  ├─ Permission Settings (supports button-level and data-level permissions)
│  ├─ Form Permissions (controls field disabling and hiding)
│  ├─ Department Management
├─ Logs
│  ├─ Document Operation Records
├─ Workflow
│  ├─ Workflow Designer
└─ Other Modules
   └─ More features under development...