🚀 PromptLens - AI Analytics & Chat Platform

What It Does

PromptLens is a comprehensive AI-powered analytics and chat platform that helps organizations understand, analyze, and optimize their AI interactions. It's essentially a "business intelligence tool for AI conversations" that provides deep insights into how people interact with AI systems.

Core Functionality

🧠 1. AI Chat Analytics

• Conversation Tracking: Records all AI prompts and responses with full metadata
• Vector Similarity Search: Uses embeddings to find similar conversations and responses
• Keyword Selection: Algorithmically selected relevant keyword for fast context and accurate embeddings
• Smart Caching: Reuses similar responses to reduce API costs and improve response times
• Multi-LLM Support: Works with OpenAI, Anthropic Claude, and XAI (Grok) models
• Reduce Repeated Requests Automatically generate markdown files for commonly asked questions to reduce costs and environmental impact

📊 2. Natural Language Data Visualization

• Query in Plain English: Ask questions like "Show me daily prompt volume trends" or "What are the most common user questions?"
• Novel NLP Approach: Iteratively increases the amount of context the LLM has until accurate categorization is achieved
• Automatic Chart Generation: Converts natural language queries into interactive charts (line, bar, pie charts)
• Time Series Analysis: Supports different granularities (daily, hourly, 30-minute, 15-minute intervals)
• Real-time Insights: Provides instant visualizations of your AI usage patterns

🔍 3. Advanced Analytics Dashboard

• Usage Metrics: Track prompts, responses, users, and engagement over time
• Performance Monitoring: Response times, token usage, and model performance
• User Behavior Analysis: Understand how different users interact with AI
• Cost Optimization: Identify opportunities to reduce API costs through caching

Why It's Useful

�� For Organizations

• Cost Optimization: Reduce AI API costs by 30-50% through intelligent caching
• Quality Assurance: Monitor AI response quality and consistency, helping to improve system prompts, documentation, etc.
• Usage Insights: Understand which AI features are most valuable
• Performance Monitoring: Track response times and identify bottlenecks

�� For Data Teams

• AI Analytics: Get detailed insights into AI usage patterns
• Custom Visualizations: Create charts and dashboards from natural language queries
• Data Export: Export conversation data for further analysis
• A/B Testing: Compare different AI models and prompts

👥 For End Users

• Better AI Experience: Faster responses through caching
• Conversation History: Never lose important AI conversations
• Multi-Model Access: Use the best AI model for each task
• Intuitive Interface: Natural language queries for complex analytics

How It Was Built

🏗️ Architecture Overview

PromptLens uses a modern, scalable architecture with three main components:

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   Frontend      │    │   Backend API   │    │   Database      │
│   (Next.js)     │◄──►│   (FastAPI)     │◄──►│   (Supabase)    │
│                 │    │                 │    │                 │
│ • Dashboard     │    │ • LLM Services  │    │ • PostgreSQL   │
│ • Chat UI       │    │ • Vector Search │    │ • Vector DB    │
│ • Analytics     │    │ • Caching       │    │ • Auth         │
└─────────────────┘    └─────────────────┘    └─────────────────┘

🛠️ Technology Stack

Frontend (Next.js 14)

• Framework: Next.js with App Router
• Styling: Tailwind CSS + shadcn/ui components
• Charts: Recharts for data visualization
• State Management: React hooks and context
• Authentication: Supabase Auth

Backend (FastAPI)

• Framework: FastAPI with async/await
• LLM Integration: OpenAI, Anthropic, XAI APIs
• Vector Search: OpenAI embeddings + cosine similarity
• Caching: Intelligent response caching system
• Authentication: JWT + API key management

Database (Supabase)

• Primary DB: PostgreSQL with vector extensions
• Vector Storage: pgvector for similarity search
• Authentication: Built-in user management
• Real-time: WebSocket subscriptions for live updates

🔧 Key Technical Features

1. Vector Similarity Search

# Uses OpenAI embeddings to find similar conversations
embedding = await openai.embeddings.create(
    model="text-embedding-3-small",
    input=prompt
)

2. Intelligent Caching

• Semantic Matching: Finds similar prompts using vector similarity
• Cost Reduction: Reuses responses for similar queries
• Quality Control: Only caches high-quality responses

3. Natural Language Query Processing

// Converts "Show me daily trends" into structured data

const result = await agent.run("Show me daily trends");

4. Multi-LLM Support

• Unified Interface: Same API for all AI providers
• Model Selection: Choose the best model for each task
• Fallback Handling: Graceful degradation if models fail

📈 Scalability & Performance

• Horizontal Scaling: Stateless FastAPI backend
• Database Optimization: Indexed vector searches
• Caching Strategy: Multi-layer caching (memory + database)
• CDN Integration: Static assets served via CDN
• Real-time Updates: WebSocket connections for live data

🔐 Security & Privacy

• API Key Management: Secure storage of LLM API keys
• User Authentication: Supabase Auth with JWT tokens
• Data Encryption: All data encrypted in transit and at rest
• Access Control: Role-based permissions for different user types

Deployment

• Frontend: Deployed on Vercel (Next.js)
• Backend: Deployed on Heroku (FastAPI)
• Database: Supabase (managed PostgreSQL)
• Monitoring: Built-in logging and error tracking

This architecture makes PromptLens a powerful, scalable platform for AI analytics that can grow with organizations while providing immediate value through cost savings and insights.