🔌 pyllmlib

pyllmlib is a lightweight, provider-agnostic Python package that gives you one simple interface to work with multiple Large Language Model (LLM) APIs.

Stop juggling different SDKs and client libraries — whether it’s OpenAI GPT, Google Gemini, Mistral AI, or Groq, you write your code once and switch providers in seconds.

🎯 Ideal for developers who want to:

• Experiment with different LLMs quickly
• Build multi-provider AI applications
• Avoid vendor lock-in with a consistent API

✨ Features

• 🔌 Unified API — A single generate() function for all providers
• 🌐 Multi-Provider Support — OpenAI, Gemini, Mistral, Groq (with more coming soon)
• 🧠 Consistent Message Format — Same request style across providers
• 🔐 Flexible Config — Use env vars, inline setup, or config files
• 📦 Minimal Dependencies — Only needs requests
• 🔄 Quick Provider Switching — Change models with one line
• 🛡️ Automatic Token Handling — Prevents overflows & context errors
• 📜 Role-Based Conversations — System, user, assistant messages
• 🔧 Extensible — Add your own providers with minimal code
• 🚀 No Vendor Lock-In — Swap providers without rewriting logic

📦 Installation

From PyPI (recommended):

pip install pyllmlib

From GitHub:

# Latest release
pip install git+https://github.com/yazirofi/pyllmlib.git

# Development version
git clone https://github.com/yazirofi/pyllmlib.git
cd pyllmlib
pip install -e .

Requirements:

• Python 3.7+
• requests (installed automatically)

🚀 Quick Start

from pyllmlib import config, generate

# Configure your preferred LLM
config(
    provider="openai",
    api_key="your-openai-api-key",
    model="gpt-4"
)

# Generate text
response = generate("Explain quantum computing in simple terms")
print(response)

✅ Same code works with any provider — just change the config.

⚙️ Configuration

1. Direct in Code

from pyllmlib import config

# OpenAI
config(provider="openai", api_key="sk-...", model="gpt-4")

# Google Gemini
config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")

# Mistral
config(provider="mistral", api_key="...", model="mistral-large-latest")

# Groq
config(provider="groq", api_key="gsk_...", model="mixtral-8x7b-32768")

2. Environment Variables

LLM_PROVIDER=openai
LLM_API_KEY=sk-your-openai-key
LLM_MODEL=gpt-4
LLM_BASE_URL=https://api.openai.com/v1  # Optional

from pyllmlib import config, generate

config()  # Loads from env
print(generate("What is LLM?"))

💬 Usage Examples

Text Generation

from pyllmlib import config, generate

config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")

print(generate("What is the capital of France?"))

prompt = """
Write a Python function to calculate factorial with error handling and docstring.
"""
print(generate(prompt))

Interactive Chat

from pyllmlib import config, chat, reset_chat

config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")

while q := input("Ask: "):
    print(chat(q))

reset_chat()

Style Generated Output Response

from pyllmlib import config, chat, reset_chat, style, generate

config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")

style(generate("Write the wish in one line"))  # use the **style** function instead of **print** function

while q := input("Ask: "):
    style(chat(q))  # use the **style** function instead of **print** function
    
reset_chat()

🌐 Supported Providers

✅ Currently Supported

• OpenAI — gpt-4, gpt-4-turbo, gpt-3.5-turbo
• Google Gemini — gemini-2.5-flash, gemini-1.5-flash
• Mistral AI — mistral-large-latest, mistral-small-latest
• Groq — mixtral-8x7b-32768, llama2-70b-4096, gemma-7b-it

🔜 Coming Soon

• Anthropic Claude
• Cohere
• Ollama & LM Studio (local hosting)
• Hugging Face models

🐛 Troubleshooting

• Auth Errors → Check API key format
• Model Not Found → Verify model name is correct

📊 Best Practices

# ✅ Reuse config for multiple prompts
config(provider="openai", api_key="sk-...", model="gpt-4")
for p in prompts:
    print(generate(p))

# ❌ Don’t reconfigure on every request

💡 Cost Optimization: Use gpt-3.5-turbo for simple tasks, gpt-4 for complex ones.

📚 API Reference

• config(**kwargs) → Set provider, API key, model
• generate(prompt, **kwargs) → Single text output
• generate_stream(prompt, **kwargs) → Streaming output
• chat(message) → Conversational interface
• reset_chat() → Clear conversation history

📄 License

Licensed under the MIT License – see LICENSE.

👨💻 Author

Shay Yazirofi

• GitHub: @yazirofi
• Email: yazirofi@gmail.com

⭐ If you find pyllmlib useful, please star the repo on GitHub! 📖 More docs & tutorials: Wiki