NoLimitAI

NoLimitAI is a Python library designated to route LLM requests across multiple providers (Groq, OpenRouter, Together AI, Gemini, Mistral) with built-in round-robin load balancing and automatic failover. It simplifies managing multiple AI services through a single, unified asynchronous API.

Features

• Multi-Provider Support: Seamlessly switch between Groq, OpenRouter, Together AI, Google Gemini, and Mistral AI.
• Round-Robin & Fallback: Automatically rotates through configured providers to distribute load and handles failures by retrying with the next available service.
• Async Streaming: Native async/await support with streaming responses.
• Unified Interface: Use one standard API for all providers.

Installation

Install the package from PyPI:

pip install nolimit-ai

You can view the project on PyPI here: https://pypi.org/project/nolimitai/

Usage

Here is a complete example of how to configure and use NoLimitAI.

1. Basic Setup

You need to provide API keys for the services you want to use. The library does not automatically load environment variables; you must pass them explicitly.

import asyncio
import os
from nolimitai import NolimitAI

# Optional: Load environment variables from a .env file
# from dotenv import load_dotenv
# load_dotenv()

async def main():
    # Initialize the client
    nlai = NolimitAI()

    # Configure the client with your API keys and optional parameters
    # You only need to provide keys for the services you intend to use.
    nlai.set_config(
        temperature=0.7,
        max_tokens=1024,
        top_p=0.9,
        keys={
            "groq": os.getenv("GROQ_API_KEY"),
            "openrouter": os.getenv("OPENROUTER_API_KEY"),
            "gemini_ai": os.getenv("GEMINI_API_KEY"),
            "mistral_ai": os.getenv("MISTRAL_API_KEY"),
            "together_ai": os.getenv("TOGETHER_API_KEY"),
        }
    )

    prompt = "Explain the concept of round-robin scheduling in one sentence."

    print(f"--- Asking: {prompt} ---\n")

    # Check which service is next in line (optional)
    next_service = nlai.get_next_service()
    print(f"[Next Service]: {next_service}")

    try:
        # Stream the response
        print("[Response]: ", end="", flush=True)
        async for token in nlai.chat(prompt=prompt):
            print(token, end="", flush=True)
        print("\n")
        
        # Check which service actually handled the request
        used_service = nlai.get_last_used_service()
        print(f"[Used Service]: {used_service}")

    except Exception as e:
        print(f"\nAn error occurred: {e}")

if __name__ == "__main__":
    asyncio.run(main())

2. Configuration Options

The set_config method accepts the following parameters:

• temperature (float): Sampling temperature (0.0 to 1.0).
• max_tokens (int): Maximum number of tokens to generate.
• top_p (float): Nucleus sampling parameter.
• keys (dict): A dictionary mapping provider names to API keys.

Supported Keys:

• "groq"
• "openrouter"
• "together_ai"
• "gemini_ai"
• "mistral_ai"

Important Notes

• Failover: If a service fails to connect or authorize, the router will automatically try the next configured service in the list.

Troubleshooting

• RuntimeError: NoLimitIA no ha sido configurado: This error occurs if you try to call chat() before calling set_config(). Ensure you configure the instance with at least one valid API key.
• Authentication Errors: Ensure your API keys are correct. If a key is invalid, the router will treat it as a failure and attempt to switch to another provider if available.

License

This project is licensed under the Apache-2.0 License.

Documentación en español