📋 AI-Powered RFP Management System

Automate your procurement workflow — generate RFPs from plain English, send them to vendors, and let AI compare proposals for you.

---

🚀 What This Does

Traditional RFP (Request for Proposal) processes are slow, manual, and error-prone. This system replaces that workflow with an AI-powered pipeline:

1. Describe your need in plain English → AI generates a structured RFP
2. Select vendors from your list → System sends RFPs via email automatically
3. Vendors reply to the email → System parses their proposals with AI
4. Compare proposals side-by-side with AI scoring on price, timeline, and fit

---

✨ Features

Feature	Description
🗣️ Natural Language Input	"I need 20 laptops for engineering, budget 50k" → structured RFP
📧 Email Automation	Send RFPs and receive proposals via Gmail integration
🤖 AI Proposal Parsing	Automatically extracts price, timeline, and terms from vendor emails
⚖️ AI Comparison	Side-by-side scoring and recommendation across all proposals
🏢 Vendor Management	Full CRUD for your vendor database
📖 Swagger Docs	All API endpoints documented and testable at /api

---

🛠️ Tech Stack

• Frontend: React 19, TypeScript, Vite, TailwindCSS
• Backend: NestJS, Express, TypeScript
• Database: MongoDB + Mongoose
• AI: Google Gemini (@google/generative-ai)
• Email: Nodemailer (SMTP) + imap-simple (IMAP) for Gmail
• API Docs: Swagger / OpenAPI

---

⚡ Quick Start

Prerequisites

• Node.js v18+
• MongoDB (local or Atlas)
• Google AI Studio API key
• Gmail account with App Password enabled

Backend Setup

cd backend
npm install
cp .env.example .env   # Fill in MongoDB URI, Gemini API key, Gmail credentials
npm run start:dev

Frontend Setup

cd frontend
npm install
cp .env.example .env   # Set VITE_API_URL=http://localhost:3000 if needed
npm run dev

Backend API: http://localhost:3000
Frontend: http://localhost:5173
Swagger Docs: http://localhost:3000/api

---

🏗️ Project Structure

AI-Powered-RFP-Management-System/
├── frontend/               # React + TypeScript + Vite UI
│   └── src/
├── backend/                # NestJS API
│   ├── src/
│   │   ├── rfps/           # RFP module (create, send, manage)
│   │   ├── vendors/        # Vendor CRUD module
│   │   ├── proposals/      # Proposal parsing & comparison
│   │   └── ai/             # Gemini integration service
│   └── .env.example
└── AI-Powered RFP Management System.pdf   # System design document

---

📡 API Overview

RFPs

Method	Endpoint	Description
POST	/rfps	Create RFP from natural language
GET	/rfps	List all RFPs
GET	/rfps/:id	Get single RFP
POST	/rfps/:id/send	Send RFP to vendors

Proposals

Method	Endpoint	Description
GET	/proposals/rfp/:rfpId	Get all proposals for an RFP
POST	/proposals/check-emails	Trigger email inbox check
GET	/proposals/compare/:rfpId	Run AI comparison

Example — Create RFP:

POST /rfps
{ "request": "I need 20 laptops for engineering team, budget 50k" }

---

🧩 Key Technical Decisions

• MongoDB chosen for flexibility with AI-generated, semi-structured JSON data
• Structured output prompting — LLM prompted for "ONLY JSON" with fallback markdown-stripping for reliability
• Email IMAP polling — system checks inbox for vendor replies and auto-parses them
• NestJS modules keep RFP, Vendor, Proposal, and AI logic cleanly separated

---

📄 License

MIT — feel free to use, fork, and build on this.

---

Built by Mohit Bansal · LinkedIn · askaide.in