The AI-powered developer lifecycle platform that understands, cleans, migrates, tests, triages, heals, and secures your projects — end to end.
CodeSentinel DP is a unified AI-powered platform that solves 13 real enterprise problems in one continuous workflow. It demonstrates how AI can autonomously handle the entire recovery and modernization lifecycle of a legacy codebase and its data, with advanced features for voice interaction, automated documentation, incident debugging, code review coaching, and stakeholder communication.
Voice-enabled AI assistant with dual AI engines - choose between Groq (fast) or DialoGPT transformers (deep learning). Features:
- Dual AI Modes:
- 🚀 Groq Mode: Lightning-fast responses with Llama 3.3 70B (default)
- 🧠 Transformer Mode: Deep learning with DialoGPT + BERT embeddings for conversational AI
- Voice Input: Click microphone and speak naturally - auto-sends after 2 seconds of silence
- Voice Output: Toggle voice responses on/off for hands-free interaction
- Enhanced RAG: Better semantic search with BERT (all-mpnet-base-v2) embeddings
- Wikipedia Integration: Enrich knowledge base with Wikipedia articles on-demand
- Real-time Transcription: See your speech converted to text in real-time
- Browser-based Speech: Uses Web Speech API (Chrome/Edge) - no server-side speech processing needed
AI Models:
- Groq: Llama 3.3 70B (fast, cloud-based)
- Transformers: DialoGPT-medium + BERT (local, conversational)
- Embeddings: all-mpnet-base-v2 (better than MiniLM)
Perfect for: Hands-free code review, rapid onboarding, accessibility, multitasking developers, and offline AI
Translates technical code into plain English for executives, product managers, and business stakeholders. Adapts explanation depth to audience.
AI Engine: Groq API (Llama 3.3 70B) | Voice: Web Speech API (TTS)
Auto-generate unit tests, integration tests, and coverage reports for codebases using CodeLlama.
AI Engine: Groq API (CodeLlama 34B) | Analysis: AST parsing + AI generation
Analyze GitHub issues, assign severity, and link to similar past incidents using RAG.
AI Engine: Groq API (Llama 3.3 70B) | RAG: ChromaDB + LangChain
Paste error logs and get instant root cause analysis with execution flow tracing and actionable fixes. Finds similar historical incidents.
AI Engine: Groq API (Llama 3.3 70B) | RAG: ChromaDB + LangChain
Analyzes sentiment in code review comments and rewrites harsh feedback into constructive, actionable suggestions. Generates AI-powered reviews.
AI Engine: Groq API (Llama 3.3 70B) | Sentiment: TextBlob + VADER
Automated cleaning, handling missing values, and numerical encoding for ML readiness.
AI Engine: Pandas + Scikit-Learn | Analysis: Statistical profiling
Analyze and normalize SQL database schemas with AI-powered normalization recommendations.
AI Engine: Frontend-only logic | Analysis: SQL parsing + normalization rules
Speak your questions and get instant voice responses about your codebase. Perfect for hands-free code exploration and rapid onboarding.
AI Engine: Groq API (Llama 3.3 70B) | Voice: Web Speech API (Browser-based)
Analyze legacy monoliths and generate modern microservices with Docker + OpenAPI.
AI Engine: Groq API (Llama 3.3 70B) | Analysis: AST parsing + microservice generation
Monitor CI/CD runs, detect failures, parse logs via LLM, and auto-generate fix patches.
AI Engine: Groq API (Llama 3.3 70B) | Analysis: Log parsing + patch generation
Analyze preprocessed datasets and project context to receive AI-powered algorithm recommendations.
AI Engine: Groq API (Llama 3.3 70B) | Analysis: Dataset profiling + ML recommendations
Automatically generate documentation and unit tests when code changes. AI-powered classification and routing.
AI Engine: Groq API (CodeLlama 34B) | Analysis: AST parsing + AI generation
RAG ChatBot → Code Explainer → QA Agent
↓ ↓ ↓
Triage → Incident Debugger → Review Coach
↓ ↓ ↓
Dataset Cleaner → DB Optimizer → Repo V-Assist
Code Porter → CI/CD Healer → Model Suggester → Doc & Test Pipeline
- Backend: Python 3.11 + FastAPI
- AI Engines:
- Groq API: Llama 3.3 70B, CodeLlama 34B (Fast & Free)
- Transformers: DialoGPT-medium, BERT (PyTorch-based)
- Speech: Web Speech API (Browser-based STT/TTS)
- Sentiment: TextBlob, VADER Sentiment
- Data Science: Pandas, Scikit-Learn, Joblib
- Vector DB: ChromaDB (local)
- RAG: LangChain + Sentence Transformers (all-mpnet-base-v2)
- Deep Learning: PyTorch 2.11+, Transformers 4.35+
- Knowledge: Wikipedia API integration
- Frontend: React 18 + Vite, TailwindCSS, Monaco Editor, Lucide Icons
- Database: SQLite (local)
- Python 3.11+
- Node.js 18+
- Git installed
- Clone the repository:
git clone https://github.com/sarthar3/CodeSentinel-DP.git
cd CodeSentinel-DP- Copy environment variables:
cp .env.example .env- Edit
.envand add your API keys:
GROQ_API_KEY=your_groq_key
# Optional for IBM Watson features
WATSON_API_KEY=your_watson_key
WATSON_URL=your_watson_url
# Optional for Google Search / Gemini features
GOOGLE_API_KEY=your_google_key
- Install dependencies:
# Backend
cd backend
pip install -r requirements.txt
# Frontend
cd ../frontend
npm install --legacy-peer-deps --forcecd backend
python -m uvicorn backend.api.main:app --reloadAvailable at http://localhost:8000
cd frontend
npm run devAvailable at http://localhost:5173
- 13-Feature Platform: Complete developer lifecycle from documentation to deployment
- RAG-Powered: Hallucination-free responses with source citations
- Multi-Language: Supports Python, JavaScript, TypeScript, Java, Go, and more
- Real-time Analysis: Instant code analysis and recommendations
- Voice Interaction: Hands-free code review and onboarding
- Public Interface: No authentication required for rapid local development
- Persistent History: All interactions stored locally in SQLite
- AI-Driven Logic: Each feature uses tailored prompts and RAG
- Streaming Responses: Real-time AI output for better UX
- Extensible Architecture: Easy to add new features and integrations
MIT License