What is Unified Knowledge System

The Unified Knowledge Management System is a comprehensive platform designed to integrate multiple knowledge sources into a unified, searchable knowledge base, enhancing information retrieval through advanced search capabilities.

Use cases

Use cases include academic research where multiple data sources need to be aggregated, corporate environments for managing internal knowledge bases, and AI applications that require context-aware information retrieval.

How to use

Users can set up the system by following the installation instructions provided in the README. Once installed, users can input various data sources, and the system will process and unify this information for easy access.

Key features

Key features include advanced search capabilities, integration of diverse knowledge sources, a six-layer architecture for efficient knowledge management, and support for vector search and knowledge graphs.

Where to use

The system can be utilized in various fields such as education, research, corporate knowledge management, and any domain that requires efficient information retrieval from multiple sources.

🧠 Unified Knowledge Management System

A comprehensive system for integrating multiple knowledge sources into a unified, searchable knowledge base with advanced retrieval capabilities.

Table of Contents

• 🧠 Unified Knowledge Management System
  • Table of Contents
  • Overview
  • Architecture
  • Components
    • Web Content Acquisition
    • Knowledge Processing
    • Knowledge Storage
    • MCP Server Layer
    • Integration Layer
    • Client Layer
  • Setup Instructions
    • Prerequisites
    • Installation
  • Performance Benchmarks
  • Cost Comparison
  • Implementation Timeline
  • Development
  • License

Overview

The Unified Knowledge Management System is designed to aggregate, process, and unify knowledge from various sources, making it accessible through standardized interfaces. This system bridges the gap between different knowledge repositories, creating a seamless experience for both users and AI assistants.

By combining vector search, knowledge graphs, and traditional document storage, our system provides comprehensive knowledge retrieval with high relevance and context awareness.

Architecture

The system is built around a six-layer architecture, each handling specific aspects of knowledge management:

graph TD
    A[Web Content Acquisition Layer] --> B[Knowledge Processing Layer]
    B --> C[Knowledge Storage Layer]
    C --> D[MCP Server Layer]
    D --> E[Integration Layer]
    E --> F[Client Layer]
    
    subgraph "Web Content Acquisition"
        A1[DevDocs - Free] 
        A2[Firecrawl - $16/month]
    end
    
    subgraph "Knowledge Processing"
        B1[Text Chunking]
        B2[Vector Embedding]
        B3[Entity Extraction]
    end
    
    subgraph "Knowledge Storage"
        C1[Structured Docs Store]
        C2[Qdrant Vector DB]
        C3[Knowledge Graph DB]
        C4[Obsidian Vault]
    end
    
    subgraph "MCP Servers"
        D1[DevDocs MCP]
        D2[Firecrawl MCP]
        D3[Qdrant MCP]
        D4[Knowledge Graph MCP]
    end
    
    subgraph "Integration"
        E1[Unified Search Engine]
        E2[Supergateway]
    end
    
    subgraph "Clients"
        F1[Claude Desktop]
        F2[Cursor]
        F3[Roo Code]
    end
    
    A --> A1 & A2
    B --> B1 & B2 & B3
    C --> C1 & C2 & C3 & C4
    D --> D1 & D2 & D3 & D4
    E --> E1 & E2
    F --> F1 & F2 & F3

For a detailed architecture description, see ARCHITECTURE.md.

Components

Web Content Acquisition

• DevDocs: Free and open-source documentation crawler with 364 GitHub stars, capable of processing up to 1000 pages/minute
• Firecrawl: Commercial web crawler ($16/month) handling general web content at approximately 20 pages/minute

Knowledge Processing

• Text Chunking: Divides documents into 512 token chunks with 50 token overlap
• Vector Embedding: Creates semantic representations using Sentence Transformers (all-MiniLM-L6-v2)
• Entity Extraction: Identifies entities and relationships for knowledge graph construction

Knowledge Storage

• Structured Documents Store: Preserves original documents with metadata in JSON/Markdown format
• Qdrant Vector Database: High-performance vector similarity search (626 QPS at 99.5% recall)
• Knowledge Graph Database: Stores entities and their relationships for graph-based queries
• Obsidian Vault: Manages personal knowledge with bidirectional linking

MCP Server Layer

• DevDocs MCP: Exposes technical documentation via stdio transport
• Firecrawl MCP: Exposes web content via HTTP+SSE transport
• Qdrant MCP: Provides vector search capabilities via stdio transport
• Knowledge Graph MCP: Enables graph-based queries via stdio transport

Integration Layer

• Unified Search Engine: Combines and ranks results from multiple knowledge sources
• Supergateway: Handles protocol conversion (stdio↔SSE) and client connection

Client Layer

• Claude Desktop: AI assistant with unified knowledge access
• Cursor: AI-enhanced code editor with knowledge integration
• Roo Code: AI coding assistant leveraging unified knowledge

Setup Instructions

Prerequisites

• Node.js 18+
• Python 3.9+
• Docker and Docker Compose
• Git

Installation

1. Clone the repository

   git clone https://github.com/BjornMelin/unified-knowledge-system.git
   cd unified-knowledge-system
   

2. Set up DevDocs

   cd mcp-servers/devdocs
   npm install
   ./setup.sh
   

3. Set up Firecrawl

   cd ../firecrawl
   npm install
   cp config.example.json config.json
   # Edit config.json with your API key
   

4. Deploy Qdrant

   cd ../qdrant
   docker compose up -d
   npm install
   

5. Set up Knowledge Graph

   cd ../knowledge-graph
   npm install
   cp config.example.json config.json
   

6. Configure Obsidian Integration

   cd ../../obsidian
   npm install
   ./setup.sh
   

7. Set up Unified Search

   cd ../integration/unified-search
   npm install
   cp config.example.json config.json
   

8. Configure Supergateway

   cd ../supergateway
   npm install
   cp config.example.json config.json
   

9. Configure Clients

   cd ../../client-configs
   ./setup-clients.sh
   

For detailed setup instructions for each component, see the README.md file in each component directory.

Performance Benchmarks

Cost Comparison

Implementation Timeline

Development

See CONTRIBUTING.md for contribution guidelines and WORKFLOW.md for our Git workflow standards.

License

This project is licensed under the MIT License - see the LICENSE file for details.