AI Chatbot Platform

A multi-tenant embeddable AI Chatbot Platform (SaaS) with RAG capabilities, built with React and FastAPI.

Features

• 🏢 Multi-tenant Architecture: Each organization has isolated data and chatbots
• 🤖 Multiple Chatbots per Tenant: Create and manage multiple chatbots
• 📚 Knowledge Ingestion: URLs, PDFs, PowerPoints, text, and Q&A pairs
• 🔍 RAG Pipeline: Chunk → Embed → Vector Store → Retrieve → Generate
• 💬 Two Embed Modes: Chat bubble script and iframe embed
• 📊 Admin Dashboard: Analytics, conversation history, and management
• 🔐 Secure Authentication: JWT + refresh tokens with RBAC (admin/editor)
• ⚡ Rate Limits & Billing: Usage tracking and quota management
• 🎨 Customizable Appearance: Theme colors, bot names, avatars

Tech Stack

Backend

• FastAPI (Python 3.11+)
• PostgreSQL with SQLAlchemy + Alembic
• ChromaDB/FAISS for vector storage
• Google Gemini for embeddings + LLM

Frontend

• React 18 with TypeScript
• Vite for build tooling
• TailwindCSS for styling
• React Query for data fetching
• Zustand for state management
• Recharts for analytics charts

Project Structure

├── backend/
│   ├── app/
│   │   ├── api/           # API routes
│   │   ├── core/          # Config, security, logging
│   │   ├── db/            # Database models
│   │   ├── schemas/       # Pydantic schemas
│   │   ├── services/      # Business logic
│   │   └── main.py        # FastAPI app
│   ├── alembic/           # Database migrations
│   ├── static/            # Widget scripts
│   └── requirements.txt
│
├── frontend/
│   ├── src/
│   │   ├── components/    # React components
│   │   ├── pages/         # Page components
│   │   ├── lib/           # API client
│   │   ├── store/         # Zustand stores
│   │   └── types/         # TypeScript types
│   └── package.json
│
└── README.md

Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+
• PostgreSQL 14+
• Google Gemini API key

Backend Setup

1. Create and activate virtual environment:

cd backend
python -m venv venv
# Windows
venv\Scripts\activate
# Linux/Mac
source venv/bin/activate

2. Install dependencies:

pip install -r requirements.txt

3. Configure environment:

# Copy example and edit with your values
cp env.example .env

Required environment variables:

DATABASE_URL=postgresql+asyncpg://user:pass@localhost:5432/chatbot_platform
SECRET_KEY=your-32-char-secret-key
JWT_SECRET_KEY=your-32-char-jwt-secret
GEMINI_API_KEY=your-gemini-api-key

4. Create database:

# In PostgreSQL
CREATE DATABASE chatbot_platform;

5. Run migrations:

alembic upgrade head

6. Start the server:

uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

Frontend Setup

1. Install dependencies:

cd frontend
npm install

2. Start development server:

npm run dev

The frontend runs at http://localhost:3000 and proxies API requests to the backend.

Usage

1. Register an Account

Visit http://localhost:3000/register to create an account and organization.

2. Create a Chatbot

1. Go to Chatbots → Create Chatbot
2. Configure name, system prompt, appearance
3. Save the chatbot

3. Add Knowledge

1. Open chatbot → Knowledge Base
2. Add URLs, upload PDFs/PPTs, or add Q&A pairs
3. Wait for ingestion to complete

4. Embed the Chatbot

1. Go to chatbot settings → Get Embed Code
2. Copy the script and add to your website:

<!-- Chat Bubble -->
<script>
  (function() {
    var script = document.createElement('script');
    script.src = 'http://your-domain.com/widget.js';
    script.setAttribute('data-chatbot-id', 'YOUR_CHATBOT_ID');
    script.async = true;
    document.body.appendChild(script);
  })();
</script>

Or use iframe:

<iframe
  src="http://your-domain.com/embed/chat/YOUR_CHATBOT_ID"
  width="400"
  height="600"
  frameborder="0"
></iframe>

API Documentation

Once the backend is running, visit:

• Swagger UI: http://localhost:8000/api/docs
• ReDoc: http://localhost:8000/api/redoc

Key Endpoints

Endpoint	Description
POST /api/auth/register	Register new user & tenant
POST /api/auth/login	Login and get tokens
GET /api/chatbots	List chatbots
POST /api/chatbots	Create chatbot
POST /api/chatbots/{id}/knowledge/url	Ingest URL
POST /api/chatbots/{id}/knowledge/file	Upload PDF/PPT
POST /api/embed/chat/{id}	Chat endpoint for widget
GET /api/admin/dashboard	Dashboard stats

Vector Store Adapters

The platform supports multiple vector databases through an adapter interface:

# In backend/app/core/config.py
VECTOR_DB_TYPE=chroma  # or 'faiss'

To add a new adapter (e.g., Pinecone):

1. Create backend/app/services/vector_stores/pinecone.py
2. Implement the VectorStoreAdapter interface
3. Add to factory in vector_store.py

Deployment

Backend (Render/Heroku)

1. Set environment variables
2. Use gunicorn with uvicorn workers:

gunicorn app.main:app -w 4 -k uvicorn.workers.UvicornWorker

Frontend (Vercel/Netlify)

1. Build the frontend:

npm run build

2. Deploy the dist folder

3. Configure API URL in production

Database

Use a managed PostgreSQL service:

• Render PostgreSQL
• Heroku Postgres
• Supabase
• AWS RDS

Configuration Options

Chatbot Settings

Setting	Description	Default
system_prompt	Instructions for the AI	"You are a helpful assistant."
temperature	Response creativity (0-2)	0.7
max_tokens	Max response length	1000
theme_color	Widget color	#6366f1
allowed_domains	CORS whitelist	[] (all)

Tenant Limits

Plan	Message Limit	Features
Free	1,000/month	Basic
Pro	10,000/month	Priority support
Enterprise	Custom	SLA, dedicated

Security

• Passwords hashed with bcrypt
• JWT tokens with short expiry
• Refresh token rotation
• Rate limiting per tenant/session
• Domain allowlisting for embeds
• SQL injection protection via ORM

License

MIT License - see LICENSE file for details.