https://github.com/dsaikiran01/e-commerce-chat-assistant
AI Shopping Assistant built with LangGraph.js, MongoDB, and React. An intelligent agent that searches, reasons, and adapts like a real sales associate—powered by vector search, multi-turn memory, and dynamic tool use.
https://github.com/dsaikiran01/e-commerce-chat-assistant
agentic-ai gemini langgraph langgraph-js mern vector-search
Last synced: 10 months ago
JSON representation
AI Shopping Assistant built with LangGraph.js, MongoDB, and React. An intelligent agent that searches, reasons, and adapts like a real sales associate—powered by vector search, multi-turn memory, and dynamic tool use.
- Host: GitHub
- URL: https://github.com/dsaikiran01/e-commerce-chat-assistant
- Owner: dsaikiran01
- Created: 2025-08-23T10:48:13.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-08-23T11:42:08.000Z (11 months ago)
- Last Synced: 2025-08-23T17:26:45.394Z (11 months ago)
- Topics: agentic-ai, gemini, langgraph, langgraph-js, mern, vector-search
- Language: TypeScript
- Homepage:
- Size: 297 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 🤖 AI Shopping Assistant with LangGraph.js & MongoDB
**An AI-powered shopping assistant that interacts, reasons, and adapts like a real sales associate.**
---
## 🎯 Overview
This project is an advanced **AI Agent** for e-commerce, designed using an **agentic architecture**. Unlike traditional chatbots, this agent:
- 🧠 **Thinks**: Determines the best course of action for a user's request
- 🔍 **Acts**: Searches product catalogs using semantic vector embeddings
- 🔄 **Adapts**: Falls back to keyword-based search if needed
- 💬 **Remembers**: Maintains context over long conversations
---
## 📽️ Demo
https://github.com/user-attachments/assets/1043b1d9-15c2-48fd-a4b5-d89daf41c0f3
---
## 🤔 Why Agentic AI?
| 🤖 Traditional Chatbot | 🧠 Agentic AI System |
| ---------------------- | ----------------------------- |
| Static replies | Autonomous decision-making |
| No real tools | Real-world API + search tools |
| Single-turn dialogs | Contextual, multi-turn memory |
| No learning | Adaptive and flexible |
---
## 📚 Topics
### 🧠 **Conceptual Knowledge**
- Agentic AI architectures
- LangGraph.js orchestration flows
- Vector-based search using MongoDB Atlas
- Multi-turn conversation design
### 🛠️ **Technical Skills**
- Node.js backend with REST APIs
- AI integrations (OpenAI, Gemini)
- React frontend with live AI chat
- Seeding and embedding product data
---
## 🚀 Prerequisites
### 📦 Software Requirements
- [Node.js](https://nodejs.org/) (v18+)
- Git & Terminal
### 🔐 API Access
- [Google AI API Key](https://aistudio.google.com/app/apikey)
- [MongoDB Atlas URI](https://www.mongodb.com/cloud/atlas)
---
## ⚡ Quick Start
### 1️⃣ Clone & Install Dependencies
```bash
git clone https://github.com/dsaikiran01/E-commerce-Chat-Assistant.git
cd E-commerce-Chat-Assistant/server
npm install
````
### 2️⃣ Setup Environment Variables
Create a `.env` file in the `/server` directory:
```env
PORT=8000
GOOGLE_API_KEY=your_google_api_key
MONGODB_ATLAS_URI=your_mongodb_uri
```
### 3️⃣ Seed the Database
```bash
npm run seed
```
🔍 What This Does
* Generates synthetic product data
* Embeds item descriptions for vector search
* Stores everything in MongoDB
### 4️⃣ Start the Backend
```bash
npm run dev
```
The backend will be running on `http://localhost:8000`.
---
## 🧪 Try the Agent
### Start a Conversation
```bash
curl -X POST http://localhost:8000/chat \
-H "Content-Type: application/json" \
-d '{"message": "I need a modern coffee table"}'
```
### Continue the Conversation
```bash
curl -X POST http://localhost:8000/chat/ \
-H "Content-Type: application/json" \
-d '{"message": "What's the price range?"}'
```
---
## 🖥️ Frontend Setup
### 5️⃣ Launch React Client
```bash
cd ../client
npm install
npm run start
```
Visit `http://localhost:3000` to start chatting with your AI shopping assistant.
---
## 🧩 Architecture Diagram
```mermaid
graph TD
A[User] --> B[LangGraph Agent]
B --> C{Decision Engine}
C -->|Search| D[Vector Search Tool]
C -->|Direct| E[Language Model Response]
D --> F[MongoDB Atlas]
F --> G[Search Results]
G --> E
E --> H[Final Response]
```
---
## 🌟 Features
| 🧠 Smart Decisions | 🔍 Search Intelligence | 💬 Conversational Memory |
| ------------------- | --------------------------- | ------------------------ |
| Dynamic tool usage | Vector + fallback search | Multi-turn support |
| Context-aware logic | Real-time inventory lookups | Persistent threads |
| Adaptive strategies | Semantic understanding | Human-like tone |
---
## 📡 API Reference
| Method | Endpoint | Description |
| ------ | ----------------- | ------------------------ |
| `GET` | `/` | Health check |
| `POST` | `/chat` | Start a new conversation |
| `POST` | `/chat/:threadId` | Continue a conversation |