🚀 GPT Fusion

The Python toolkit that makes AI integration effortless

GPT Fusion is a comprehensive Python library designed to streamline AI-assisted application development. It ships a real LLM client for OpenAI-compatible chat completions (OpenAI, Groq, a local Ollama server, anything speaking the same API), plus the text processing, web scraping, and interactive demo tooling it's had all along.

🎯 Why Choose GPT Fusion?

• 🤖 Real LLM Client: OpenAI-compatible chat completions, works with OpenAI, Groq, local Ollama, or anything else on that API shape
• ⚡ Zero Setup Friction: Install and start coding in seconds
• 🛡️ Production Ready: Built-in security, performance optimizations, and error handling
• 🔌 Modular Design: Use only what you need with optional dependencies
• 📚 Rich Examples: Complete demo projects including Unity 3D games and auth systems
• 🧪 Battle Tested: 80+ tests with 89%+ coverage and CI/CD pipeline

📦 Install from PyPI • 🌐 Live Documentation • 🎮 Try Live Demo

🚀 Quick Start

📦 Installation

Quick Install:

pip install gpt-fusion

Full Installation (all features):

pip install "gpt-fusion[all]"

Python Version Support:

• ✅ Python 3.10+
• ✅ Python 3.11 (Recommended)
• ✅ Python 3.12

⚡ Quick Start Example

import gpt_fusion

# 🔤 Smart text processing
text = "The quick brown fox jumps over the lazy dog"
print(f"Words: {gpt_fusion.word_count(text)}")
print(f"Reversed: {gpt_fusion.reverse_words(text)}")
print(f"Is palindrome: {gpt_fusion.is_palindrome('racecar')}")

# 📊 Powerful data analysis
data = gpt_fusion.load_numbers_from_csv('data/numbers.csv')
print(f"Average: {gpt_fusion.average_from_csv('data/numbers.csv', use_streaming=True)}")

# 🌐 Easy web scraping (with built-in security)
headlines = gpt_fusion.scrape("https://news.ycombinator.com", ".titleline > a")
print(f"Found {len(headlines)} headlines")

# 🚀 Generate small, real starter projects in seconds
print(gpt_fusion.create_csv_app('my-analytics-dashboard', with_api=True))
print(gpt_fusion.create_tailwind_ui('my-modern-webapp', dark_mode=True))

Output:

Words: 9
Reversed: dog lazy the over jumps fox brown quick The
Is palindrome: True
Average: 3.0
Found 30 headlines
my-analytics-dashboard
my-modern-webapp

🤖 LLM Integration

pip install "gpt-fusion[llm]"
export OPENAI_API_KEY=sk-...

from gpt_fusion import ask, LLMClient

# One-liner for a single question
reply = ask("Explain recursion in one sentence.")

# Or reuse a client across a multi-turn conversation
client = LLMClient(model="gpt-4o-mini")
reply = client.chat([
    {"role": "system", "content": "Answer in a single sentence."},
    {"role": "user", "content": "What's a closure?"},
])
client.close()

Points base_url anywhere that speaks the OpenAI chat completions API without changing your code. Verified working against Groq's free tier:

client = LLMClient(
    base_url="https://api.groq.com/openai/v1",
    model="llama-3.1-8b-instant",
    api_key=os.environ["GROQ_API_KEY"],
)

The same pattern works for a local Ollama server (base_url="http://localhost:11434/v1") or any other OpenAI-compatible endpoint.

🎛️ Optional Feature Sets

Choose the components you need:

# 🤖 LLM chat completions client
pip install "gpt-fusion[llm]"

# 🌐 Web scraping & HTTP clients
pip install "gpt-fusion[web]"

# 🚀 FastAPI backend with auto-docs
pip install "gpt-fusion[backend]"

# 🐦 Social media integration
pip install "gpt-fusion[twitter]"

# 🛠️ Asset optimization & building
pip install "gpt-fusion[build]"

# 🧪 Development tools
pip install "gpt-fusion[dev]"

# 🎯 Everything included
pip install "gpt-fusion[all]"

✨ Features

🤖 LLM Client

Chat completions for OpenAI-compatible APIs - real requests, real error handling, no vendor lock-in.

# Install: pip install "gpt-fusion[llm]"
from gpt_fusion import ask, LLMClient

ask("Summarize the plot of Hamlet in two sentences.")

client = LLMClient()  # reads OPENAI_API_KEY from the environment
client.chat("Hello!", temperature=0.2)

Raises ConfigurationError if no API key is available, and APIError if the request fails or the response comes back in an unexpected shape - both importable from gpt_fusion.

🐍 Python Utilities

Core text processing, math helpers, and CSV analysis tools.

import gpt_fusion

# Text processing
gpt_fusion.word_count("Hello world")  # 2
gpt_fusion.reverse_words("Hello world")  # "world Hello"
gpt_fusion.is_palindrome("racecar")  # True

# Math & CSV
gpt_fusion.average_from_csv("data.csv")
gpt_fusion.median_from_csv("data.csv")

🌐 Web Scraping

Simple web scraping utilities with BeautifulSoup integration.

# Install: pip install "gpt-fusion[web]"
import gpt_fusion

html = gpt_fusion.scrape("https://example.com")
# Returns clean text content

🚀 FastAPI Backend

Ready-to-deploy API server with auto-generated docs.

# Install: pip install "gpt-fusion[backend]"
import uvicorn
import gpt_fusion

# Start server
uvicorn.run(gpt_fusion.backend_app, port=8000)

🐦 Twitter Integration

Twitter bot utilities with OAuth support.

# Install: pip install "gpt-fusion[twitter]"
# Reads TWITTER_API_KEY, TWITTER_API_SECRET, TWITTER_ACCESS_TOKEN, and
# TWITTER_ACCESS_SECRET from the environment if not passed explicitly.
from gpt_fusion import TwitterBot

bot = TwitterBot()
bot.post_tweet("Hello from GPT Fusion!")

🎮 Interactive Demos

🔐 Enhanced Auth UI Kit

Modern, secure authentication system with comprehensive security features:

• 🛡️ Rate limiting & input sanitization
• 🎨 Beautiful glass-effect UI with dark mode
• ♿ WCAG 2.1 AA accessibility compliance
• 🔍 Real-time password strength validation
• 📱 Fully responsive design

Try it: Enhanced Demo | Basic Version | Test Suite

🎯 Unity 3D Game Engine Integration

Complete game architecture demonstrating modern Unity patterns:

• ⚡ Event-driven systems (no Update() polling)
• 🏊 Object pooling for performance
• 🎛️ Scriptable Object configuration
• 🖥️ Modern UI with smooth animations
• 🏗️ Interface-based architecture

Explore: Modern Scripts | Setup Guide

📊 Data Analysis Playground

High-performance CSV processing with streaming support for large datasets:

• ⚡ Memory-efficient streaming for large files
• 📈 Statistical analysis (mean, median, percentiles)
• 🔒 Built-in security (path traversal protection)
• 📋 Sample datasets included

$ python examples/tutorial.py
Values: [1.0, 2.0, 3.0, 4.0, 5.0]
Average: 3.0
Median: 3.0

Try: Tutorial Script | Sample Data

🛠️ Project Generator

Small, real starter kits - not empty stubs, each command below actually runs and produces working code:

# 📊 CSV demo script, plus a FastAPI wrapper over the same data
gpt-fusion create_csv_app my-analytics --with-api
# -> my-analytics/{app.py, numbers.csv, api.py}

# 🎨 Tailwind + Firebase auth UI (--dark-mode for the glass-effect variant)
gpt-fusion create_tailwind_ui my-webapp --dark-mode
# -> my-webapp/{index.html, app.js}

# 🚀 Both combined: a frontend/ + backend/ FastAPI app
gpt-fusion create_fullstack_app my-saas --auth --database
# -> my-saas/frontend/{index.html, app.js}
# -> my-saas/backend/{app.py, numbers.csv}

--auth adds a minimal HMAC-signed-token login flow (POST /login with demo/demo123, then a bearer token on every other route) and --database swaps reading the CSV live for a small SQLite-backed store - both demo-grade and clearly commented as such in the generated code, not production-hardened. Neither flag adds a new dependency beyond what gpt-fusion[backend] already needs.

(python -m gpt_fusion <command> ... works the same way as gpt-fusion <command> ... if you'd rather not rely on the installed console script.)

🚀 API & Deployment

💻 Local Development

Start the development server:

pip install "gpt-fusion[backend]"
uvicorn gpt_fusion.backend:app --reload --port 8000

Interactive Features:

• 📝 Swagger UI: http://localhost:8000/docs
• 🔧 ReDoc: http://localhost:8000/redoc
• 📊 Health Check: http://localhost:8000/health

🌐 API Endpoints

🌐 Cloud Deployment

🚀 Deploy to Render (Recommended)

# render.yaml is included in the repo
git push origin main  # Auto-deploys via GitHub integration

🟣 Deploy to Heroku

# Procfile is included
heroku create my-gpt-fusion-app
git push heroku main

🐳 Deploy with Docker

FROM python:3.11-slim
WORKDIR /app
COPY . .
RUN pip install "gpt-fusion[backend]"
EXPOSE 8000
CMD ["uvicorn", "gpt_fusion.backend:app", "--host", "0.0.0.0", "--port", "8000"]

🛠️ Troubleshooting

Common Issues

❌ Installation fails on Python 3.9

# GPT Fusion requires Python 3.10+
pyenv install 3.11.0
pyenv local 3.11.0
pip install gpt-fusion

❌ Import errors with optional dependencies

# Install specific feature sets
pip install "gpt-fusion[web]"  # for scraping
pip install "gpt-fusion[backend]"  # for FastAPI

❌ CSV files not loading

# Ensure CSV has 'value' column header
import gpt_fusion
data = gpt_fusion.load_numbers_from_csv('data.csv', use_streaming=True)

🔍 Still having issues?

• 📞 GitHub Issues
• 📚 Documentation

🤝 Contributing

🛠️ Development Setup

git clone https://github.com/costasford/gpt-fusion.git
cd gpt-fusion
pip install "gpt-fusion[dev]"  # Installs dev dependencies
pip install -e .  # Editable install
pre-commit install  # Git hooks for quality

🧪 Testing & Quality

# Run the full test suite (80+ tests)
pytest

# Check coverage (currently 89%+)
pytest --cov=src/gpt_fusion --cov-report=html

# Code formatting and linting
black .
flake8 .

# Run all quality checks
python scripts/run_checks.py

Project Structure

src/gpt_fusion/     # Main package
├── core.py         # Basic utilities  
├── llm.py          # LLM chat completions client (optional)
├── text_utils.py   # Text processing
├── analysis.py     # CSV/data tools
├── web_scraper.py  # Web scraping (optional)
├── backend.py      # FastAPI server (optional)
├── twitter_bot.py  # Twitter integration (optional)
└── starter_kits.py # Project templates

tests/              # Comprehensive test suite
docs/               # Jekyll documentation
examples/           # Usage examples

Links

📖 GitHub Repository • 📦 PyPI Package • 🐛 Report Issues • 📄 MIT License

GPT Fusion - Practical demos of human-AI collaboration. Built with ❤️ and Python.