A SaaS platform that transforms any website into a deployable AI-powered chatbot. Enter a URL, and the system crawls the website, builds a RAG knowledge base using Spring AI + PGVector, and generates a complete chatbot package ready for deployment.
- Smart Web Crawling — Jsoup-powered crawler follows internal links + sitemap.xml
- Content Processing — Cleans HTML, removes boilerplate, chunks text semantically
- Automatic AI Failover — Self-healing RAG pipeline (OpenAI → Anthropic → Google)
- User-Provided API Keys — Users can use their own OpenAI/Anthropic keys for project isolation
- Switchable Engines — Toggle between AI providers instantly via UI or environment variables
- Free Key Guide — Detailed guide on getting free API keys for all models
- Code Generation — Auto-generates a FastAPI chatbot server + embeddable widget
- ZIP Packaging — Download a ready-to-deploy chatbot with docs & deployment scripts
- Modern Dashboard — Dark theme Next.js UI with real-time progress tracking
- JWT Authentication — Secure user accounts with isolated project data
- Chat Preview — Test your chatbot directly in the dashboard
┌─────────────────┐ ┌─────────────────────────────────────────┐
│ Next.js UI │ │ Spring Boot Backend │
│ (Port 3000) │─── REST ─│ │
└─────────────────┘ │ ┌────────┐ ┌─────────┐ ┌──────────┐ │
│ │Crawler │→│Chunker │→│Spring AI │ │
│ │(Jsoup) │ │ │ │Embed+RAG │ │
│ └────────┘ └─────────┘ └──┬───────┘ │
│ │ │
│ ┌────────────────────────┐ │ │
│ │ Auth + Projects │ │ │
│ │ Code Gen + ZIP │ │ │
│ └────────────────────────┘ │ │
└───────────────────────────────┼──────────┘
│
┌───────▼──────┐
│ PostgreSQL │
│ + PGVector │
│ (Port 5432) │
└──────────────┘
Key: Everything runs in a single JVM. No separate Python service, no ChromaDB — just Spring Boot + PostgreSQL.
| Component | Technology |
|---|---|
| Backend API | Spring Boot 3.2, Java 17, Spring Security, JPA |
| AI / RAG | Spring AI 1.0, OpenAI, Anthropic (Claude), Google (Gemini) |
| Vector Store | PGVector (PostgreSQL extension via Spring AI) |
| Web Crawling | Jsoup 1.17 |
| Frontend | Next.js 14, TypeScript, Tailwind CSS |
| State Mgmt | Redux Toolkit (auth) + TanStack Query v5 (server state) |
| Forms | React Hook Form + Zod schema validation |
| Database | PostgreSQL 16 + PGVector |
| Auth | JWT (jjwt 0.12) — separate secrets per service |
| Deployment | Docker, Docker Compose |
 |
 |
|---|---|
| Project Dashboard | Multi-Model Configuration |
 |
 |
|---|---|
| Real-time Crawling | Context-Aware Chat Preview |
Follow these steps to run the complete project locally on your terminal.
Use Docker to run the database with the vector extension.
# This starts the database on port 5433 (to avoid conflicts with local Postgres)
docker compose up -d postgresNavigate to the backend folder and run with Maven.
cd backend
mvn spring-boot:runNote: Ensure your backend/.env is configured with your OPENAI_API_KEY.
Navigate to the frontend folder and start the dev server.
cd frontend
npm run devThe UI will be available at http://localhost:3000.
Use these commands to verify your data inside the running Docker container:
- List Tables:
docker exec -it instantchatbot-postgres psql -U postgres -d instantchatbot -c "\dt"
- View Vector Data:
docker exec -it instantchatbot-postgres psql -U postgres -d instantchatbot -c "SELECT * FROM vector_store LIMIT 5;"
- Interactive Shell:
docker exec -it instantchatbot-postgres psql -U postgres -d instantchatbot
| Method | Endpoint | Description |
|---|---|---|
| POST | /api/auth/register |
Register new user |
| POST | /api/auth/login |
Login, returns JWT |
| GET | /api/auth/me |
Get current user |
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/projects |
List user's projects |
| POST | /api/projects |
Create new project |
| GET | /api/projects/{id} |
Get project details |
| Method | Endpoint | Description |
|---|---|---|
| POST | /api/projects/{id}/crawl |
Start crawling + RAG pipeline |
| GET | /api/projects/{id}/status |
Get processing status |
| POST | /api/projects/{id}/chat |
Send chat message (Spring AI RAG) |
| GET | /api/projects/{id}/download |
Download chatbot ZIP |
The entire AI pipeline runs inside the Spring Boot JVM:
URL Input → WebCrawlerService (Jsoup)
→ ContentCleanerService (HTML cleanup)
→ TextChunkerService (semantic splitting)
→ EmbeddingService (Spring AI → OpenAI → PGVector)
→ RagService (Spring AI ChatClient → OpenAI GPT-4o-mini)
Services in com.instantchatbot.service.ai:
| Service | Responsibility |
|---|---|
WebCrawlerService |
Crawls website using Jsoup, follows links + sitemap |
ContentCleanerService |
Strips boilerplate, extracts main content |
TextChunkerService |
Splits text into semantic chunks with overlap |
EmbeddingService |
Generates embeddings, stores in PGVector, searches |
RagService |
Retrieves context + generates answers with multi-model failover |
AiModelOrchestrator |
Manages global vs. user-provided AI models and keys |
CrawlPipelineService |
Orchestrates the full async pipeline |
When you download the ZIP, you get a standalone FastAPI chatbot:
chatbot-server/
├── main.py # FastAPI server with /chat endpoint
├── requirements.txt # Python dependencies
├── .env.example # Configuration template
├── Dockerfile # Container deployment
├── README.md # Setup & deployment guide
├── deploy.sh # Auto-deployment script
└── widget/
└── chatbot-widget.js # Embeddable chat widget
- Separate Secret Keys — Backend (
JWT_SECRET) and frontend (NEXTAUTH_SECRET) each have independent secrets - JWT Authentication — Stateless token-based auth signed with the backend secret
- Password Hashing — BCrypt encryption
- Data Isolation — PGVector metadata filtering per project
- Zod Validation — Client-side schema validation on all forms
- URL Validation — Server-side input sanitization
- CORS Configuration — Configurable allowed origins
PENDING → CRAWLING → PROCESSING → EMBEDDING → READY
↘ FAILED
MIT License
Built with ❤️ using Spring Boot, Spring AI, PGVector, Next.js, and OpenAI