KbNet Overview

KBNet is an interactive learning platform that creates dynamic, AI-powered knowledge maps for users to explore topics in an engaging and intuitive way. The project uses MindsDB for AI capabilities and follows a modern web architecture.

Try KbNet

• Online Trial: Experience KbNet at https://kbnet.diybuilds.tech/

• Self-Host Pack: Download the latest self-host pack from our GitHub releases and read setup instructions in the self-host/README.md file.

• Demo Video: Watch KbNet in action on YouTube

• Project Article: Learn about the development and vision behind KbNet in our detailed article

Core Features

1. Knowledge Map Exploration

• Users can create personalized knowledge maps starting from any topic
• Interactive navigation using swipe gestures (UP/DOWN/LEFT/RIGHT)
• Three types of topic relationships:
  • DEEP: More detailed exploration of current topic
  • RELATED: Connected but different concepts
  • SIMILAR: Alternative approaches or perspectives

2. AI-Powered Content Generation

• Uses MindsDB integration with Gemini-2.0-flash model
• Dynamically generates:
  • Topic summaries
  • Related concepts
  • Learning paths
• Knowledge base (KB_NAME) for contextual information

3. Progress Tracking

• User stats and achievements system
• XP-based progression:
  • Start map: 50 XP
  • Visit node: 5 XP
  • Return to node: 2 XP
  • Daily streak bonus: 20 XP
• Badges for accomplishments
• Streak tracking system

4. Map Summaries

• AI-generated summaries of exploration paths
• Journal-style narratives of learning journeys
• 24-hour cooldown between summary generations
• Status tracking (PENDING/IN_PROGRESS/COMPLETED/FAILED)

Technical Architecture

Frontend

• Built with Next.js
• Located in platform
• Features:
  • Interactive map visualization
  • Progress tracking dashboard
  • User achievements display
  • Summary generation interface

Backend

• Node.js server in server
• Key controllers:
  • Map generation and navigation
  • User statistics tracking
  • Achievement system
  • Summary generation

Database Schema

Main tables:

• Maps: User's learning journeys
• Nodes: Individual topics
• Navigation Steps: User's exploration path
• Node Relationships: Topic connections
• Map Summaries: Generated learning narratives

AI Integration

• MindsDB for AI processing
• Multiple data sources:
  • Wikipedia
  • HackerNews
  • YouTube
• Custom prompts for different node types
• Specialized summary generation agent

Developer Features

• Turborepo monorepo setup
• TypeScript throughout
• ESLint and Prettier configuration
• Remote caching support
• Shared component library
• Development and production environment configurations

Notable Technical Decisions

1. Use of MindsDB for AI processing
2. Monorepo architecture with shared packages
3. Real-time WebSocket updates for user interactions
4. Separation of platform and server applications
5. TypeScript for type safety across the codebase

This project represents a sophisticated implementation of AI-powered educational technology, focusing on interactive and personalized learning experiences.

Setup Instructions

Prerequisites

• Node.js (v18 or later)
• Docker