AI-Powered SDLC Automation using LangGraph

An AI-powered Software Development Life Cycle (SDLC) automation platform built using LangGraph, FastAPI, OpenRouter, and Mem0.

This project simulates a complete software development lifecycle where multiple AI agents collaborate to transform software requirements into deployable project artifacts while performing iterative reviews, security analysis, testing, and deployment recommendations.

---

Features

• AI Requirement Analysis
• Agile User Story Generation
• Product Owner Review & Revision
• Software Design Document Generation
• AI Design Review
• Project Skeleton Code Generation
• AI Code Review
• Security Review
• Automated Security Fix Suggestions
• Test Case Generation
• AI Test Case Review
• QA Testing
• Deployment Recommendation
• Long-Term Memory using Mem0
• LangGraph State Machine Workflow
• REST API using FastAPI
• OpenRouter LLM Integration

---

Workflow

Requirements
      │
      ▼
User Story Generation
      │
      ▼
Product Owner Review
      │
      ├──────────────┐
      ▼              │
Revise Stories ◄─────┘
      │
      ▼
Design Generation
      │
      ▼
Design Review
      │
      ├──────────────┐
      ▼              │
Revise Design ◄──────┘
      │
      ▼
Code Generation
      │
      ▼
AI Code Review
      │
      ├──────────────┐
      ▼              │
Revise Code ◄────────┘
      │
      ▼
Security Review
      │
      ├──────────────┐
      ▼              │
Fix Security ◄───────┘
      │
      ▼
Generate Test Cases
      │
      ▼
Test Case Review
      │
      ├──────────────┐
      ▼              │
Revise Test Cases ◄──┘
      │
      ▼
QA Testing
      │
      ▼
Deployment Recommendation

---

AI Agents

• Requirement Analyst
• Product Owner
• Software Architect
• Design Reviewer
• Backend Developer
• Code Reviewer
• Security Engineer
• Security Fix Agent
• Test Case Generator
• QA Engineer
• Deployment Advisor

---

Tech Stack

Backend

• Python
• FastAPI

AI Framework

• LangGraph
• LangChain
• OpenRouter
• Mem0

LLM

• OpenRouter Models

Validation

• Pydantic

Environment Management

• python-dotenv

---

Long-Term Memory (Mem0)

The project integrates Mem0 to provide persistent memory across workflow executions.

Current implementation includes:

• Storing project requirements
• Retrieving relevant memories before User Story Generation
• Retrieving relevant memories before Design Generation

This enables the AI agents to maintain contextual awareness across multiple software development sessions.

---

API

Run Workflow

POST

/run

Example Request

{
    "requirements": "Build a Bank Management System"
}

---

Example Output

The workflow generates:

• Agile User Stories
• Software Design Document
• Project Skeleton
• Code Review Feedback
• Security Review Report
• Generated Test Cases
• QA Testing Report
• Deployment Recommendation

---

Deployment Recommendation

The workflow recommends deployment platforms such as:

• Vercel
• Render
• Railway
• AWS
• Azure

---

Future Improvements

• LangGraph Checkpointer
• Thread-based Conversations
• Resume Workflow
• Human-in-the-loop Interrupts
• Multi-user Support
• Docker Support
• CI/CD Integration
• Code Execution Sandbox
• Automated Documentation Generation

---

Why This Project?

This project demonstrates how multiple AI agents can collaboratively automate different phases of the Software Development Life Cycle using a state-driven workflow.

It showcases:

• Agentic AI
• Multi-Agent Systems
• LangGraph State Machines
• Prompt Engineering
• Long-Term Memory with Mem0
• AI-Powered Code Review
• Security Analysis
• Automated Testing
• End-to-End SDLC Automation

---

Author

Rishu Raj