⚡ DevIntel AI Enterprise v2

Autonomous Developer Productivity & RAG Knowledge Platform

DevIntel AI is an enterprise-grade AI assistant that transforms documentation websites, GitHub repositories, and code bases into an interactive Knowledge Graph, visual architecture flowcharts, automated PR reviewer, and token-by-token streaming RAG search engine.

Live Demo • API Documentation • Deployment Guide • Benchmarks

---

📌 Repository Description & Topics

Description: Autonomous AI Developer Productivity Platform featuring multi-page web scraping, hybrid BM25 + Cosine similarity vector search, token streaming RAG chat, AI Coding Agent, automated PR code reviewer, repository health metrics, physics-simulated Knowledge Graphs, and Model Context Protocol (MCP) tool connectors.

Topics / Tags: developer-tools • rag • vector-database • fastapi • nextjs • langchain • firecrawl • gemini-api • ai-agent • code-refactoring • pull-request-reviewer • knowledge-graph • mermaid-diagrams • prometheus-metrics

---

🏗️ System Architecture & Data Pipeline

graph TD
    Client["🌐 Next.js 14 App Router (Frontend)"] --> Gateway["⚡ FastAPI REST & SSE Gateway"]
    
    subgraph Data Ingestion & Scraping
        Gateway --> Scraper["🕸️ Firecrawl / Parallel Scraper"]
        Gateway --> GHAPI["🐙 GitHub API Profile & Repo Extractor"]
        Scraper --> Chunker["✂️ Markdown Sliding Window Chunker"]
        GHAPI --> Chunker
    end

    subgraph Vector Retrieval Engine
        Chunker --> Embedder["🧬 Gemini Embedder (text-embedding-004)"]
        Embedder --> VectorDB["💾 ChromaDB Cosine Store"]
        Gateway --> HybridRAG["🔍 Hybrid BM25 + Cosine Search"]
        VectorDB --> HybridRAG
    end

    subgraph Multi-LLM & AI Swarm
        HybridRAG --> LLM["🤖 Multi-LLM Engine (Gemini, GPT-4o, Claude 3.5, DeepSeek R1)"]
        LLM --> Stream["💬 SSE Token Streaming Chat & Citations"]
        LLM --> CodingAgent["💻 AI Coding Agent (Multi-File Refactorer)"]
        LLM --> PRReviewer["🛡️ Pull Request Reviewer AI"]
    end

Loading

---

🌟 Key Enterprise v2 Features

Feature	Description	Status
🌐 Ultra-Fast Parallel Scraping	Ingests documentation & GitHub profiles in ~2-3s via ThreadPoolExecutor.	✅
💬 SSE Token Streaming RAG	Real-time typewriter responses with grounded URL citations [1], [2].	✅
💻 Autonomous AI Coding Agent	Generates multi-file patch diffs and line-by-line refactoring instructions.	✅
🔍 Pull Request Review AI	Evaluates PR code diffs for Security, Bugs, Style, Complexity, & Docs.	✅
📈 Repository Health & Timeline	Health metrics rating (Overall 94/100) and commit evolution visualizer.	✅
🕸️ Physics Knowledge Graph	Interactive HTML5 Canvas physics simulation with entity drag & filter pills.	✅
📐 System Architecture Diagrams	Auto-generates project-tailored Mermaid.js flowcharts with source code export.	✅
🤖 Specialized AI Agent Swarm	Security Reviewer (OWASP audit), Architecture Reviewer, & Interview Coach.	✅
🔑 SaaS API Keys & Audit Ledger	Secret API keys (sk_live_...), quota limits, and immutable audit trail.	✅
🔌 MCP Connectors Client	Integrates with GitHub, Slack, Notion, Linear, Jira, & Google Drive.	✅

---

🚀 Quick Start Guide

Prerequisites

• Python 3.10+
• Node.js 18+

1. Clone Repository & Setup Backend

git clone https://github.com/Rishisharma029/DevIntel-AI.git
cd DevIntel-AI

# Install Python dependencies
pip install -r requirements.txt

# Start FastAPI server
python -m uvicorn backend.main:app --host 127.0.0.1 --port 8000

2. Setup Frontend Studio

cd frontend

# Install Node dependencies
npm install

# Start Next.js dev server
npm run dev

Open http://localhost:3000 in your browser.

---

🧪 Automated Testing & Benchmarks

Run Pytest test suite:

pytest tests/test_backend.py
# Result: 4 passed in 0.85s

Run Performance Benchmarks:

python benchmarks/benchmark_suite.py
# Benchmark Report published to BENCHMARKS.md

Metric	Result	Target
Embedding Throughput	2,840.5 tokens/sec	> 1,500 t/s
Vector Retrieval Latency	0.12 ms	< 5.0 ms
Crawl Processing Speed	18.4 pages/sec	> 10.0 p/s
Concurrent Query Capacity	450 req/sec	> 200 req/s

---

🐳 Docker Deployment

Run full platform via Docker Compose:

docker-compose -f docker/docker-compose.yml up --build

---

📜 License & Community

• License: Released under the MIT License.
• Code of Conduct: Review our CODE_OF_CONDUCT.md.
• Security Policy: Read our SECURITY.md.
• Contributing: See CONTRIBUTING.md.

---

_{Built with ❤️ by Rishi Sharma & DevIntel AI Engineering Team}