EcommForAll is a modern, AI-first e-commerce platform that blends traditional shopping with intelligent automation. Built with Spring AI and Google Vertex AI, it features conversational assistance, semantic search, and automated content generation to deliver a cutting-edge shopping experience.
✨ [Link to Live Demo (Coming Soon)] ✨
The fastest way to run the entire stack (Frontend, Backend, PostgreSQL + pgvector):
git clone https://github.com/ainan-ahmed/EcommForAll.git
cd EcommForAll
docker-compose up --buildFor manual setup without Docker:
- Prerequisites: Java 21, Node.js 18+, Maven, PostgreSQL + pgvector.
- Database: Install
pgvector(e.g.,apt install postgresql-15-pgvector) and create a database namedecommforall. - Backend:
cd backend ./mvnw clean install ./mvnw spring-boot:run - Frontend:
cd frontend npm install npm run dev
- Frontend:
http://localhost:5173| Backend API:http://localhost:8080| Swagger:/swagger-ui.html
For cloud features, follow the GCP Setup Guide.
| Layer | Primary Technologies | Key Libraries & Frameworks |
|---|---|---|
| Frontend | React, TypeScript, Vite | Mantine UI, TanStack Router/Query, Zustand, Zod |
| Backend | Java 21, Spring Boot 3.4 | Spring Data JPA, Spring Security, Spring AI, MapStruct |
| AI/ML | Google Vertex AI, pgvector | Gemini 3 Flash, Text Embedding 005, HNSW Indexing |
| Storage | PostgreSQL, AWS S3 | Hibernate, Amazon SDK |
| DevOps | Docker, GitHub Actions | Docker Compose, Reviewdog, Maven |
- Auth & Roles: JWT-based secure login with Admin, Seller, and User roles.
- Product Engine: Complex variants, attributes, and inventory management.
- Brand/Category: Hierarchical category system and brand CRUD.
- Shopping Flow: Fully functional Cart, Wishlist, Checkout, and Reviews.
- Intelligent Chatbot: Conversational shopping assistant with memory and tool calling.
- Semantic Search: Natural language search powered by 768-dim vector embeddings.
- AI Content Gen: Automated product descriptions with customizable tone and style.
- Recommendations: Vector-similarity based "Similar Products" engine.
| Component | Detail |
|---|---|
| Primary Model | Gemini 3 Flash (via Vertex AI) |
| Embeddings | Google Text Embedding 005 (768 Dimensions) |
| Vector Store | pgvector (PostgreSQL) with HNSW Indexing |
| Memory | JDBC-backed sliding window (Default: 10 messages) |
| Distance Metric | Cosine Similarity |
For API payloads and implementation details, see AI Service Documentation.
- ✅ Phase 1 (Complete): Auth, Product Engine, Vector Search, AI Chatbot, Reviews, Checkout.
- 🚧 Phase 2 (In Progress): Admin Dashboard, Vision AI Search, Voice Assistant integration.
- 📝 Project Structure - Codebase organization.
- ⚙️ CI/CD Workflows - Automated checks and formatting.
- 🤖 OpenCode Agents - CLI agents (
@gh-bot) and dev commands. - 🤝 Contributing - Guidelines for local development.
🤖 Built with an AI-First Philosophy by Ainan ❤️
Special thanks to: [Spring AI], [Vertex AI], [Mantine], [TanStack], and the [pgvector] community.