How asharibali/edugpt works
Main Components Identification
a) Frontend Layer:
• Next.js application (indicated by the app/ directory)
• UI Components (components/ directory)
• Authentication System (OCID integration)
b) Backend Services:
• FlowiseAI (LangChain orchestration)
• OpenAI Services (GPT-4, Whisper API, Moderation API)
• Blockscout Integration
• Google Custom Search API
• LangSmith API
┌─ Client Layer ─────────────────────┐
• Next.js Frontend │
• TailwindCSS Styling │
• TypeScript Components │
• ┬───────────────────┘
• ↓
┌─ Authentication Layer ─────────────┐
• Open Campus ID SDK │
• User Management │
• ┬───────────────────┘
• ↓
┌─ API Gateway Layer ────────────────┐
• API Routes │
• Request Handling │
• ┬────────────┬──────────┘
• ↓ ↓
┌─ AI Services ─┐ ┌─ Blockchain ────┐
• FlowiseAI │ │ - Blockscout │
• OpenAI │ │ API Gateway │
• LangChain │ │ - Custom Tools │
• ┘ └───────────────┘
Data Flow Representation
Following key interactions:
• User → Frontend → Authentication
• Frontend → AI Services
• AI Services → Blockchain Services
• AI Services → External APIs
Additional Elements
Important components:
• Knowledge Base integration (knowledgebase/ directory)
• Prompt Engineering flow (prompt-engineering/ directory)
• External Service Dependencies
• Rate Limiting Systems
• Error Handling (app/error directory)
Documentation References
References to:
• API Documentation
• User Query Examples
• System Limitations
• Authentication Flow
For implementation, I recommend using tools like:
• PlantUML
• Lucidchart