proxai 0.3.2

ProxAI is a lightweight abstraction layer for foundational AI model connections.

ProxAI

⚡️ One Interface for Every AI Model! ⚡️

Official Site • Overview • Docs • Community

Philosophy

ProxAI simplifies AI integration by providing a unified interface for connecting to multiple AI providers.

import proxai as px

result = px.generate_text(
    'Hello model! What is 23 + 45?',
    provider_model=('claude', 'sonnet'))
print(result)

result = px.generate_text(
    'Hello model! What is 23 + 45?',
    provider_model=('openai', 'gpt-4o'))
print(result)

• ⭐️ Simple Unified API: Pythonic, easy, intuitive, and unified API for all AI models connection.
• 💡 All Major AI Providers: Including Gemini, OpenAI, Claude, Grok, DeepSeek, Cohere, Mistral, and more.
• 🚀 Always Up-to-Date: Let ProxAI handle all new released model integrations, you just list and pick.
• 💻 Model-agnostic AI development: Write python code without thinking which AI provider and model you will use.

Features

• ⏰ Ready to Start: Takes 2 minutes to connect and get responses from all major models.
• 🔍 Pick and Switch: Experiment with different models to find the best fit.
• ⛑️ Robust Error Handling: Comprehensive error handling for API failures.
• 💾 Caching: Speed up responses and reduce costs with built-in query and model caching.
• 💰 Cost Estimation: Basic tracking of your estimated API call costs breakdown.
• 🥂 Integrations Status: See the status of all your integrations at a glance.
• 📊 Analytics: Total token usage, performance metrics, and more.

Quick Start

Check out our Quick Start Guide for a step-by-step guide on how to get started with ProxAI.

1. Install ProxAI:

   pip install proxai
   

2. Set API Keys:

   Export your AI provider API keys as environment variables:

   export OPENAI_API_KEY="your-openai-key"
   export GEMINI_API_KEY="your-openai-key"
   export ANTHROPIC_API_KEY="your-anthropic-key"
   # Add other provider keys as needed
   

   or, one API key rule them all! ProxConnect🚀

   export PROXAI_API_KEY="your-proxai-key"
   

   See Provider Integrations page.

3. Basic Usage:

   import proxai as px
   
   # Write model agnostic function
   def get_meaning_of_universe():
     response = px.generate_text('What is the meaning of universe?')
     print(response)
   
   # List available models
   provider_models = px.models.list_models()
   
   # Generate response for each model
   for provider_model in provider_models:
     px.set_model(provider_model)
     print(f"Testing {provider_model} model")
     get_meaning_of_universe()
   

📚 Documentation

For full details on installation, all features, and advanced usage, please visit our Overview and Documentation.

📈 ProxDash (Optional Dashboard)

Enhance your ProxAI experience with ProxDash, our optional monitoring platform for usage tracking, analytics, and experiment management. The ProxAI library works perfectly standalone.

Local development setup

After cloning, run this once so on-disk file-permission changes don't surface as spurious modifications in git status:

git config --local core.fileMode false

All files in this repo are committed with the executable bit set so individual run scripts work directly. Files arrive on your disk as 0755 after git pull. If you prefer broader perms (e.g. 0777) for your local workflow, chmod -R 777 . is safe — with core.fileMode false set, the change stays invisible to git.

🤝 Contribute & Connect

• Community: Learn how to contribute
• Discord: Join our server
• Issues: Report bugs or request features
• GitHub: https://github.com/proxai/proxai

⚖️ License

MIT License. See LICENSE for details.