MonoLLM Documentation

MonoLLM is a unified Python framework for accessing multiple Large Language Model providers through a single, consistent interface. It simplifies LLM integration by abstracting away provider-specific differences while maintaining full access to advanced features.

Note

MonoLLM v0.1.2 introduces comprehensive testing utilities, improved reasoning model support, and enhanced provider compatibility.

Key Features

Unified Interface

Access OpenAI, Anthropic, Google, Qwen, DeepSeek, and other providers through one API

Advanced Capabilities

• Streaming responses for real-time interaction

• Reasoning models with thinking steps (QwQ, o1, DeepSeek R1)

• Multi-turn conversations with context management

• Automatic model capability detection

Developer Experience

• Type-safe async/await API

• Comprehensive error handling

• Flexible configuration management

• Built-in retry mechanisms and rate limiting

Production Ready

• Proxy support for enterprise environments

• Token usage tracking and cost management

• Extensive logging and monitoring

• Comprehensive test suite

Quick Start

Installation

Install MonoLLM using pip:

pip install monollm

Or from source:

git clone https://github.com/cyborgoat/MonoLLM.git
cd MonoLLM
pip install -e .

Basic Usage

import asyncio
from monollm import UnifiedLLMClient, RequestConfig

async def main():
    async with UnifiedLLMClient() as client:
        config = RequestConfig(model="gpt-4o", temperature=0.7)
        response = await client.generate("Explain quantum computing", config)
        print(response.content)

asyncio.run(main())

Configuration

Set up your API keys:

export OPENAI_API_KEY="your-openai-key"
export ANTHROPIC_API_KEY="your-anthropic-key"
export DASHSCOPE_API_KEY="your-qwen-key"
export DEEPSEEK_API_KEY="your-deepseek-key"

Supported Providers

Provider	Models	Special Features	Status
OpenAI	GPT-4o, GPT-4o-mini, o1, o1-mini	Reasoning models, MCP	✓ Full support
Anthropic	Claude 3.5 Sonnet, Claude 3.5 Haiku	MCP integration	✓ Full support
Qwen/DashScope	QwQ-32B, Qwen3 series	Thinking steps, Chinese	✓ Full support
DeepSeek	DeepSeek V3, DeepSeek R1	Code reasoning	✓ Full support
Google	Gemini 2.0 Flash, Gemini 2.5 Pro	Multimodal (planned)	✓ Basic support
Volcengine	Doubao models	Enterprise features	✓ Basic support

Documentation Sections

User Guide

• Quick Start Guide
  • Installation
  • Configuration
  • Basic Usage
  • Command Line Interface
  • Error Handling
  • Testing Your Setup
  • Next Steps
  • Common Issues
• Configuration
  • Environment Variables
  • Configuration Files
  • Runtime Configuration
  • Advanced Configuration
  • Configuration Validation
  • Common Configuration Patterns
  • Configuration Priority
  • Troubleshooting Configuration
  • Best Practices
  • Security Considerations
• Examples
  • Basic Examples
  • Advanced Examples
  • Use Case Examples
  • Error Handling Examples
  • Performance Examples
  • Integration Examples
  • Best Practices
• Testing Guide
  • Overview
  • Test Scripts
  • Setting Up Tests
  • Test Categories
  • Understanding Test Results
  • Common Test Scenarios
  • Troubleshooting Tests
  • Custom Testing
  • Best Practices

API Reference

• Client API
  • UnifiedLLMClient
  • Usage Examples
• Data Models
  • Core Models
  • Configuration Models
  • Provider Models
  • Usage Examples
• Exceptions
  • Exception Hierarchy
  • Base Exception
  • Configuration Exceptions
  • Provider Exceptions
  • Network Exceptions
  • Validation Exceptions
  • Usage Examples
  • Best Practices

Getting Help

• GitHub Issues: Report bugs and request features at https://github.com/cyborgoat/MonoLLM/issues

• Documentation: Full documentation at https://cyborgoat.github.io/MonoLLM/

• Examples: See the examples/ directory for complete usage examples

License

MonoLLM is released under the MIT License. See the LICENSE file for details.

Indices and Tables

• Index

• Module Index

• Search Page