Project Overview

This repository contains a full-stack AI chatbot powered by IONOS AI, featuring a ReAct agent with real-time web search capabilities. The project includes:

• Backend: FastAPI service with LangGraph ReAct agent that uses Tavily for web search and integrates with both IONOS AI Model Hub (inference models) and IONOS AI Model Studio (fine-tuned models).
• Frontend: Two options:
  • Streamlit (recommended): Modern, interactive chat interface with real-time streaming support
  • Next.js (alternative): React-based web application

---

Key Features

• 🤖 Intelligent ReAct Agent: LangGraph-powered agent with tool-calling capabilities
• 🔍 Real-time Web Search: Integrated Tavily API for up-to-date information retrieval
• 🎯 Dual Model Support:
  • IONOS Model Hub (inference models with web search)
  • IONOS Studio (fine-tuned models for specialized tasks)
• ⚡ Streaming Responses: Real-time token-by-token output for Hub models
• 💬 Conversation History: Context-aware multi-turn conversations
• 🚀 Production Ready: Kubernetes deployment with CI/CD pipeline

---

Table of Contents

• Prerequisites
• Environment Variables
• Backend Setup
• Frontend Setup
  • Streamlit (Recommended)
  • Next.js (Alternative)
• Usage
• Project Structure
• Adding Fine-tuned Models
• CI/CD Deployment
• Documentation
• License

---

Prerequisites

Before you begin, ensure you have:

• Python (v3.10 or above)
• pip for Python dependencies
• Node.js (v18 or above) and npm (only if using Next.js frontend)
• An IONOS API Key for AI Model Hub access
• A Tavily API Key for web search functionality
• (Optional) IONOS Studio API key for fine-tuned models

---

Environment Variables

Backend .env

Create a .env file in the backend folder:

# Required
IONOS_API_KEY=your_ionos_api_key_here
TAVILY_API_KEY=your_tavily_api_key_here

# Optional: IONOS Studio fine-tuned models
STUDIO_API_KEY=your_studio_api_key_here
STUDIO_ORG_ID=your_studio_org_id_here
STUDIO_BASE=https://studio.ionos.de/api/v1

# Fine-tuned model UUIDs (add as many as needed)
STUDIO_MODEL_QWEN_GDPR=model_uuid_here
STUDIO_MODEL_GRANITE_GDPR=model_uuid_here
STUDIO_QWEN3_SHAREGPT=model_uuid_here

Frontend .env

Streamlit (frontends/streamlit-starter/.env):

BACKEND_URL=http://localhost:8000
IONOS_API_KEY=your_ionos_api_key_here

Next.js (frontends/next-js-starter/.env):

NEXT_PUBLIC_APP_BASE_URL=http://localhost:8000

---

Backend Setup

1. Navigate to the backend folder:

   cd backend

2. Create and activate a virtual environment:

   macOS/Linux:

   python3 -m venv .venv
   source .venv/bin/activate

   Windows:

   python -m venv .venv
   .\.venv\Scripts\Activate.ps1

3. Install dependencies:

   pip install -r requirements.txt

4. Create .env file with required variables (see Environment Variables)

5. Run the FastAPI server:

   uvicorn main:app --host 127.0.0.1 --port 8000 --reload

   The backend will be available at http://localhost:8000.

---

Frontend Setup

Streamlit (Recommended)

1. Navigate to the Streamlit folder:

   cd frontends/streamlit-starter

2. Create and activate a virtual environment:

   macOS/Linux:

   python3 -m venv .venv
   source .venv/bin/activate

   Windows:

   python -m venv .venv
   .\.venv\Scripts\Activate.ps1

3. Install dependencies:

   pip install -r requirements.txt

4. Create .env file with BACKEND_URL=http://localhost:8000 and IONOS_API_KEY

5. Run the Streamlit app:

   streamlit run app.py

   The app will open automatically at http://localhost:8501.

Next.js (Alternative)

1. Navigate to the Next.js folder:

   cd frontends/next-js-starter

2. Install dependencies:

   npm install

3. Create .env file with NEXT_PUBLIC_APP_BASE_URL=http://localhost:8000

4. Start the development server:

   npm run dev

   The app will be available at http://localhost:3000.

---

Usage

1. Start the backend (see Backend Setup)
2. Start your preferred frontend (see Frontend Setup)
3. Select a model type:
   • Inference: IONOS Hub models with web search capabilities
   • Fine-tuned: Your custom Studio models for specialized tasks
4. Choose a model from the dropdown
5. Start chatting—the agent will search the web when needed and provide informed responses

---

Project Structure

IONOS-simple-chatbot/
├── .env                        # Root environment variables (optional)
├── .env.template              # Environment template
├── README.md
├── LICENSE
├── CODE_OF_CONDUCT.md
├── docker-compose.yml         # Docker Compose configuration
├── kubernetes_config.tpl      # K8s deployment template
├── pyproject.toml            # Root Python project metadata
├── uv.lock                   # UV lockfile
│
├── backend/                   # FastAPI service
│   ├── main.py               # API endpoints and routing
│   ├── chatbot_agent.py      # LangGraph ReAct agent setup
│   ├── studio_client.py      # IONOS Studio API client
│   ├── requirements.txt      # Python dependencies
│   ├── Dockerfile            # Container image definition
│   ├── Makefile              # Build automation
│   └── .env                  # Backend environment config
│
├── frontends/
│   ├── streamlit-starter/    # Streamlit UI (recommended)
│   │   ├── app.py           # Main application
│   │   ├── requirements.txt # Python dependencies
│   │   └── Dockerfile       # Container image
│   │
│   └── next-js-starter/      # Next.js UI (alternative)
│       ├── app/             # Next.js app directory
│       │   ├── layout.tsx
│       │   ├── page.tsx
│       │   └── globals.css
│       ├── components/      # React components
│       │   ├── ChatLog.tsx
│       │   ├── Message.tsx
│       │   ├── TextInput.tsx
│       │   └── ui/         # Shadcn UI components
│       ├── lib/            # Utilities
│       ├── public/         # Static assets
│       ├── package.json
│       └── tsconfig.json
│
├── docs/                     # Documentation site (Nextra)
│   ├── pages/
│   │   ├── docs/
│   │   │   ├── getting-started/
│   │   │   ├── backend/
│   │   │   ├── frontend/
│   │   │   └── deployment/
│   │   └── index.tsx
│   ├── public/assets/       # Documentation assets
│   ├── styles/
│   ├── package.json
│   └── next.config.mjs
│
└── assets/
    └── images/              # Repository images
        ├── cover.jpg
        └── screenshot.png

---

Adding Fine-tuned Models

To add a new IONOS Studio fine-tuned model:

1. Add the model UUID to backend/.env:

   STUDIO_MODEL_YOUR_NAME=your-model-uuid-here

2. Register the model in backend/main.py:

   @app.get("/studio/models")
   async def get_studio_models():
       models = {
           "qwen-gdpr": os.getenv("STUDIO_MODEL_QWEN_GDPR"),
           "your-model-name": os.getenv("STUDIO_MODEL_YOUR_NAME"),  # Add this
       }
       return {k: v for k, v in models.items() if v}

3. Restart the backend—the model will appear in the frontend dropdown automatically

---

CI/CD: Tag-Based Build & Kubernetes Deployment

This repository includes automated CI/CD via GitHub Actions (.github/workflows/deploy.yml).

Trigger

Push a version tag matching X.Y.Z format (e.g., 1.0.0, 2.3.7):

git tag 1.0.0
git push origin 1.0.0

Workflow Steps

1. Build & Push Docker images for backend and Streamlit frontend
2. Tag images with both version number and latest
3. Deploy to IONOS Kubernetes cluster
4. Update secrets with IONOS_API_KEY and TAVILY_API_KEY
5. Validate deployment with rollout status checks

Required Secrets

Configure these in your GitHub repository settings:

• DOCKER_USERNAME / DOCKER_PASSWORD: Container registry credentials
• KUBE_CONFIG: Kubernetes cluster configuration
• IONOS_API_KEY: IONOS AI Model Hub API key
• TAVILY_API_KEY: Tavily web search API key
• STUDIO_API_KEY: (Optional) IONOS Studio API key
• STUDIO_ORG_ID: (Optional) Studio organization ID

Image Naming

Component	Version Tag	Latest Tag
Backend	<registry>/backend:1.0.0	<registry>/backend:latest
Frontend	<registry>/frontend:1.0.0	<registry>/frontend:latest

---

Documentation

Comprehensive documentation is available at the /docs directory, powered by Nextra. Topics include:

• 📖 Getting Started: Prerequisites, environment setup, installation
• 🔧 Backend: API reference, agent tools, troubleshooting
• 🎨 Frontend: Streamlit and Next.js setup guides
• 🚀 Deployment: Docker, Kubernetes, CI/CD pipelines
• 🏗️ Architecture: Project structure, model routing, agent design

To run the docs locally:

cd docs
npm install
npm run dev

Visit http://localhost:3000 to browse the documentation.

---

License

This project is released under the MIT License. Feel free to use and modify it in your own applications.

---

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

---

Support

For issues, questions, or feature requests, please open an issue on GitHub.

---

Built with ❤️ using IONOS AI, LangGraph, FastAPI, and Streamlit