🤖 Universal AI Chatbot

A powerful, multi-backend chatbot with an enhanced Gradio interface that connects to 5 different AI providers with 100+ free models.

⚠️ Important: AI model availability changes frequently. Provider APIs may add, remove, or rename models at any time. The model lists in this README are current as of December 2025 but may not reflect real-time availability.

---

✨ Features

🔌 5 AI Providers Supported

• Ollama - Run models locally (100% free, private, no API keys)
• OpenRouter - 29 free models with :free suffix
• GitHub Models - 23 free tier models for prototyping
• Groq - 19 models with ultra-fast inference (14,400 free req/day)
• Gemini - 17 Google AI models (1,500 free req/day)

🎨 Enhanced Interface

• Modern Gradio UI with custom theme
• Real-time API status indicators
• Model dropdown with 100+ options
• System prompt customization
• Temperature & max tokens controls
• Preset prompts (Code Expert, Writer, Analyst, Teacher)
• Export chat to Markdown
• Stop generation button
• Clear chat functionality

🚀 Advanced Capabilities

• Streaming responses for real-time interaction
• Conversation history support
• Empty choices safety checks (fixed streaming issues)
• Custom system prompts per conversation
• Temperature control for creativity adjustment
• Error handling with helpful suggestions

---

📋 Requirements

• Python 3.8+
• Jupyter Notebook or VS Code with Jupyter extension
• Internet connection (for cloud providers)
• API Keys (optional, depending on providers)

---

🚀 Quick Start

1️⃣ Clone Repository

git clone https://github.com/M-F-Tushar/Multi-Backend-Chatbot-with-Gradio.git
cd Multi-Backend-Chatbot-with-Gradio

2️⃣ Install Dependencies

pip install gradio openai python-dotenv google-generativeai

Or run the first code cell in the notebook:

%pip install google-generativeai openai python-dotenv gradio -q

3️⃣ Set Up API Keys

Create a .env file in the project directory:

# Copy the example template
cp .env.example .env

# Edit .env with your favorite text editor
notepad .env  # Windows
nano .env     # Linux/Mac

Add your API keys (get them from the links below). Only add keys for providers you want to use:

# OpenRouter (29 free models)
OPENROUTER_API_KEY=sk-or-v1-your-key-here

# GitHub Models (23 free models)
GITHUB_TOKEN=ghp_your-token-here

# Groq (19 free models, ultra-fast)
GROQ_API_KEY=gsk_your-key-here

# Google Gemini (17 models)
GOOGLE_API_KEY=AIzaSy-your-key-here

Note: Ollama doesn't require an API key (local only).

4️⃣ Run the Chatbot

Option A: Jupyter Notebook

jupyter notebook Chatbot.ipynb

Option B: VS Code

1. Open Chatbot.ipynb in VS Code
2. Select Python kernel
3. Run all cells (Ctrl+Shift+Enter)

Option C: JupyterLab

jupyter lab

The interface will launch at http://127.0.0.1:7860

---

🔑 Getting API Keys

🆓 Free Options (No Credit Card Required)

1. Ollama (Completely Free, Local)

# Install Ollama
# Windows/Mac: Download from https://ollama.ai
# Linux:
curl -fsSL https://ollama.com/install.sh | sh

# Pull a model
ollama pull llama3.2:1b

# Start Ollama (usually auto-starts)
ollama serve

Models Available: 13 models including Llama, Mistral, CodeLlama, Phi, Gemma, Qwen, DeepSeek

---

2. Groq (Free Tier: 14,400 requests/day)

1. Visit https://console.groq.com
2. Sign up with GitHub/Google
3. Go to "API Keys" → Create new key
4. Copy key to .env as GROQ_API_KEY

Free Tier Limits:

• 30 requests/minute
• 14,400 requests/day
• 19 models including Llama 3.3 70B, Mixtral, Gemma, Qwen

---

3. GitHub Models (Free for Prototyping)

1. Generate a Personal Access Token:
   • Go to https://github.com/settings/tokens
   • Click "Generate new token (classic)"
   • Select scopes: read:user, read:org
   • Generate and copy token
2. Add to .env as GITHUB_TOKEN

Free Tier Limits:

• 10-15 requests/minute (varies by model)
• 50-150 requests/day
• 23 models including GPT-4o, Llama, Phi, Mistral, DeepSeek

---

4. Google Gemini (Free Tier: 1,500 requests/day)

1. Visit https://aistudio.google.com/app/apikey
2. Sign in with Google account
3. Click "Create API Key"
4. Copy key to .env as GOOGLE_API_KEY

Free Tier Limits:

• 15 requests/minute
• 1,500 requests/day
• 17 models including Gemini 2.5, Gemini 3 Preview, Gemma

---

5. OpenRouter (29 Free Models with :free suffix)

1. Visit https://openrouter.ai
2. Sign up (email/Google/GitHub)
3. Go to "Keys" → Create new key
4. Copy key to .env as OPENROUTER_API_KEY

Free Tier Notes:

• Must append :free to model IDs
• ~20 requests/minute
• ~50 requests/day per model
• 29 models including Llama 3.3 70B, Gemini, DeepSeek R1, Qwen, Gemma

---

📊 Available Models (December 2025)

⚠️ Note: Model availability changes frequently. The lists below reflect models available as of December 2025. Some models may be added, removed, or renamed by providers. Always check the dropdown in the interface for the most current list.

Ollama (13 Local Models)

llama3.2:1b, llama3.2:3b, llama3.1:8b, llama3.1:70b
mistral:7b, mixtral:8x7b, codellama:7b, codellama:34b
phi3:mini, phi3:medium, gemma2:9b, qwen2.5:7b, deepseek-r1:7b

OpenRouter (29 Free Models)

meta-llama/llama-3.3-70b-instruct:free
google/gemini-2.0-flash-exp:free
tngtech/deepseek-r1t2-chimera:free
qwen/qwen3-coder:free
google/gemma-3-27b-it:free
nvidia/nemotron-nano-12b-v2-vl:free
... (24 more, see notebook for full list)

GitHub Models (23 Models)

OpenAI: gpt-4o, gpt-4o-mini, o1-mini, o1-preview
Meta Llama: Llama-3.3-70B, Llama-3.2-90B-Vision, Llama-3.1-70B
Microsoft: Phi-4, Phi-3.5-MoE
Mistral: Mistral-Large-2407, Codestral-2501
DeepSeek: DeepSeek-V3, DeepSeek-R1
... (13 more models)

Groq (19 Models)

llama-3.3-70b-versatile, llama-3.1-8b-instant
openai/gpt-oss-120b, openai/gpt-oss-20b
gemma2-9b-it, mixtral-8x7b-32768
qwen-2.5-32b, deepseek-r1-distill-llama-70b
whisper-large-v3 (audio)
... (11 more models)

Gemini (17 Models)

gemini-3-pro-preview (Preview)
gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite
gemini-2.0-flash, gemini-2.0-flash-lite
gemma-3-27b-it, gemma-3-8b-it, gemma-2-9b-it
... (8 more models)

---

🎛️ Usage Guide

Basic Usage

1. Launch the notebook and run all cells
2. Select Backend from dropdown (Ollama/OpenRouter/GitHub/Groq/Gemini)
3. Select Model from auto-populated list
4. Type your message and press Enter or click Send

Advanced Settings

Click "Advanced Settings" accordion to access:

• System Prompt - Customize AI behavior

  • Example: "You are a Python expert. Provide code with explanations."

• Temperature (0.0 - 2.0)

  • 0.0-0.5: Focused, deterministic responses
  • 0.7: Balanced (default)
  • 1.0-2.0: Creative, varied responses

• Max Tokens (0 - 4096)

  • 0: Unlimited (default)
  • 512: Short responses
  • 2048: Medium responses
  • 4096: Long responses

Preset Prompts

Quick-select specialized AI personas:

• Code Expert - For programming help
• Creative Writer - For storytelling and content
• Data Analyst - For data insights and analysis
• Teacher - For educational explanations

Export Chat

1. Click "Export Chat" accordion
2. Click "Export to Markdown"
3. Copy the formatted conversation

---

🔧 Troubleshooting

GitHub Models: "list index out of range"

✅ FIXED - Empty choices check added to streaming handler

Ollama: Connection Error

# Make sure Ollama is running
ollama serve

# Check if models are installed
ollama list

# Pull a model if needed
ollama pull llama3.2:1b

API Key Errors

1. Check .env file exists in project directory
2. Verify no extra spaces in API keys
3. Restart Jupyter kernel after editing .env
4. Run cell 2 again to reload environment

Model Not Found

• OpenRouter: Add :free suffix to free models
• GitHub: Use exact model names (e.g., Llama-3.3-70B-Instruct)
• Groq: Use model IDs from the list above
• Gemini: Some models are preview/experimental

Rate Limit Exceeded

• Free tiers have limits (see API Keys section)
• Wait a few minutes or switch to another provider
• Consider upgrading to paid tier for higher limits

---

📁 Project Structure

Multi-Backend-Chatbot-with-Gradio/
├── Chatbot.ipynb          # Main notebook with enhanced Gradio UI
├── README.md              # This documentation
├── .env.example           # Template for API keys (rename to .env and add your keys)
├── .env                   # Your API keys (DO NOT COMMIT - protected by .gitignore)
├── .gitignore             # Git configuration to protect sensitive files
└── LICENSE                # MIT License

🚀 Deployment Options

Local Development (Recommended for Testing)

jupyter notebook Chatbot.ipynb

Cloud Deployment (Optional)

Hugging Face Spaces (Free, Public)

1. Fork this repo to your GitHub account
2. Create new Hugging Face Space
3. Connect to your GitHub repo
4. Add API keys to Space secrets
5. Deploy!

Streamlit Cloud (Free)

Would require converting Gradio to Streamlit, but feasible.

Docker Deployment

Could containerize for cloud platforms (AWS, Azure, GCP, etc.)

---

🛠️ Technical Details

Architecture

• Frontend: Gradio 5.0+ with custom CSS
• Backend: OpenAI-compatible API clients
• Models: 100+ free models across 5 providers
• Streaming: Real-time token-by-token responses

Key Improvements

1. Fixed GitHub Models streaming - Added empty choices check
2. Enhanced UI - Modern theme, better layout, status indicators
3. Model management - Dynamic dropdowns with 100+ models
4. Advanced controls - System prompts, temperature, max tokens
5. Error handling - Helpful error messages with solutions

Dependencies

gradio>=5.0
openai>=1.0
python-dotenv>=1.0
google-generativeai>=0.3

---

🤝 Contributing

Found a bug? Want to add features? Here's how to contribute:

Fork & Clone

git clone https://github.com/YOUR-USERNAME/Multi-Backend-Chatbot-with-Gradio.git
cd Multi-Backend-Chatbot-with-Gradio

Create a Branch

git checkout -b feature/your-feature-name

Make Changes

• Add new features, fix bugs, improve documentation
• Test thoroughly with different models
• Update README if adding new capabilities

Commit & Push

git add .
git commit -m "Add your descriptive message"
git push origin feature/your-feature-name

Submit Pull Request

• Go to GitHub repo
• Create Pull Request from your fork
• Describe your changes

---

🌟 Ideas for Contributions

• Add more AI providers (Anthropic Claude, Mistral API, etc.)
• Vision model support (image inputs)
• Voice input/output support
• Chat history persistence (SQLite/MongoDB)
• Multi-user authentication
• Custom model fine-tuning
• API usage statistics & cost tracking
• Docker containerization
• Streamlit version
• Web server deployment guide

---

📝 License

This project is open source and available under the MIT License.

---

🙏 Acknowledgments

• Gradio - Beautiful web UI framework
• OpenAI - API standard that most providers follow
• Meta, Google, Mistral, Microsoft - Open-source models
• OpenRouter, Groq, GitHub - Free API access

---

📞 Support & Issues

Found a Bug?

1. Check existing issues
2. Create a new issue with:
   • Python version
   • Which provider failed
   • Error message (full traceback)
   • Steps to reproduce

Troubleshooting

1. Check the Troubleshooting section above
2. Verify your API keys in .env
3. Ensure all dependencies: pip install --upgrade gradio openai python-dotenv google-generativeai
4. Restart Jupyter kernel after editing .env
5. Check GitHub Discussions for community help

Feature Requests

• Open a GitHub Discussion
• Or create an Issue with [FEATURE REQUEST] label

---

🔮 Future Enhancements

• Vision model support (image inputs)
• Voice input/output
• Chat history persistence
• Multi-user support
• Custom model additions
• API usage statistics

---

Happy Chatting! 🚀

Last Updated: December 4, 2025 Repository: https://github.com/M-F-Tushar/Multi-Backend-Chatbot-with-Gradio