Cognify

Agentic Content Intelligence & Semantic Alignment Architecture

Cognify is a sophisticated, AI-driven content intelligence platform engineered to counteract engagement-optimized algorithms. Designed as an autonomous Agentic RAG (Retrieval-Augmented Generation) system, Cognify operates as a man-in-the-middle for content consumption, actively parsing, semantically embedding, and scoring digital content streams against predefined ontological user goals.

---

🏗 System Architecture & Workflows

Cognify utilizes a distributed microservice architecture comprising an edge-injected Chrome extension, a high-concurrency Node.js/Express API gateway, and a highly optimized relational persistence layer via PostgreSQL.

Data Flow & Inference Pipeline

graph TD
    User([Client]) --> |DOM Mutation Observer| Ext[Manifest V3 Extension]
    
    subgraph Browser Context
        Ext --> |Extracts Metadata & Transcripts| ExtBg[Service Worker]
        ExtBg --> |Injects Mutated DOM Elements| DOM[Shadow DOM UI]
    end
    
    subgraph Core Infrastructure
        ExtBg --> |REST / JSON Payload| API[Express API Gateway]
        API --> |JWT Auth & Rate Limiting| Auth[Middleware]
        Auth --> |In-Memory RAM Cache Pool| YTScraper[Query Aggregator Engine]
        Auth --> |Prisma ORM writes| DB[(PostgreSQL)]
    end
    
    User --> |Visualizes Aggregations| Dashboard[React + Vite Frontend]
    Dashboard <--> API

Loading

---

🗄️ Entity-Relationship Model (ERD)

The persistent data storage is handled by PostgreSQL via Prisma ORM, utilizing Foreign Keys, Cascading deletes, and highly optimized indexing for millisecond response times.

erDiagram
    USER ||--o| USER_PROFILE : owns
    USER ||--o{ CONTENT_LOG : generates
    USER ||--o{ DRIFT_SCORE : accumulates
    USER ||--o{ STREAK : maintains
    USER ||--o{ COACH_SESSION : engages
    USER ||--o{ BOOKMARK : saves
    USER ||--o{ ACHIEVEMENT : unlocks
    USER ||--o{ CIRCLE_MEMBER : joins

    COACH_SESSION ||--o{ COACH_MESSAGE : contains
    CIRCLE ||--o{ CIRCLE_MEMBER : includes

    USER {
        string id PK "UUID"
        string email UK
        datetime createdAt
    }
    
    USER_PROFILE {
        string id PK
        string userId FK
        string[] platforms "Monitored sites"
    }

    CONTENT_LOG {
        string id PK
        string url
        string title
        boolean isConscious
        int score "0-100"
        string summary
        string dominantEmotion
        int timeSpentSecs
    }
    
    DRIFT_SCORE {
        string id PK
        float score
        string trend
    }

Loading

---

🛠 Tech Stack

• Backend Infrastructure: Node.js, Express, Prisma ORM, PostgreSQL.
• Frontend Dashboard: React 18, Vite, Tailwind CSS (Strict Monochrome Apple/Linear Theme), Framer Motion, Recharts.
• Extension Integration: Chrome Manifest V3 side-panel, content scripts, and background service workers.
• Integrations: Google Gemini API, YouTube Search APIs.

---

🚀 Local Development Setup

Prerequisites

• Node.js 20+
• PostgreSQL (Local or Neon/Supabase)

1. Database & Express Backend

cd backend
npm install

# Configure your environment variables
cp .env.example .env 

# Push the schema to your Postgres instance
npx prisma db push
npx prisma generate

# Start the development server
npm run dev

2. React Frontend

cd frontend
npm install

# Start the Vite development server
npm run dev

3. Chrome Extension

1. Open Chrome and navigate to chrome://extensions/
2. Enable Developer mode in the top right.
3. Click Load unpacked and select the /cognify-extension directory in this repository.

---

🔒 Security & Performance

• Payload Optimization: The Express backend limits payloads intelligently (up to 50MB) to allow bulk extension syncing without triggering 413 Payload Too Large errors.
• Lightning Caching: The recommendation engine uses smart RAM caching and single-threaded query aggregation to prevent CPU choking on lower-tier cloud environments (like Render Free Tier).
• Type-Safety: End-to-end type safety enforced via TypeScript and Prisma Client.