llm-anygate 1.0.3

A flexible gateway for connecting and managing multiple LLM providers

LLM AnyGate

A powerful CLI tool that generates LiteLLM proxy projects from simple YAML configurations. Designed to free users from understanding the complexities of the LiteLLM library and quickly create local LLM proxies for use with various AI coding tools.

Overview

LLM AnyGate simplifies the process of setting up LiteLLM proxy servers by providing a simple command-line interface to generate complete, ready-to-run proxy projects with minimal configuration.

Key Features

🚀 Quick Setup - Create a fully configured LiteLLM proxy project with one command
📝 Simple Configuration - Use minimal YAML config instead of complex LiteLLM settings (or use defaults)
🔧 Zero Database - Generated proxies run statelessly without database requirements
🖥️ Cross-Platform - Works on Windows, macOS, and Linux with unified CLI commands
🎯 Production Ready - Generates complete project with config, environment templates, and documentation
📦 PyPI Package - Easy installation via pip from official PyPI repository

Installation

From PyPI

pip install llm-anygate

For Development (with Pixi)

# Clone the repository
git clone https://github.com/igamenovoer/llm-anygate.git
cd llm-anygate

# Initialize submodules
git submodule update --init --recursive

# Setup development environment with Pixi
pixi install
pixi shell

Quick Start

Step 1: Generate Proxy Project (Optional Configuration)

Use the CLI to generate a complete LiteLLM proxy project. The model configuration is optional:

# With default configuration (uses gpt-4o with OPENAI_API_KEY)
llm-anygate-cli create --project my-proxy

# With custom configuration file
llm-anygate-cli create \
  --project my-proxy \
  --model-config model-config.yaml \
  --port 4567 \
  --master-key "sk-my-secure-key"

If you want to use a custom model configuration, create a YAML file (model-config.yaml):

model_list:
  - model_name: gpt-4o
    litellm_params:
      model: openai/gpt-4o
      api_base: https://api.openai.com/v1
      api_key: os.environ/OPENAI_API_KEY

  - model_name: claude-3-5-sonnet
    litellm_params:
      model: anthropic/claude-3-5-sonnet-20241022
      api_key: os.environ/ANTHROPIC_API_KEY
      
  - model_name: gemini-pro
    litellm_params:
      model: gemini/gemini-pro
      api_key: os.environ/GEMINI_API_KEY

Step 2: Configure Environment

cd my-proxy

# Copy and configure environment variables
cp env.example .env
# Edit .env and add your API keys

Step 3: Start the Proxy Server

# Start using the CLI tool
llm-anygate-cli start

# Or start from within the project directory
cd my-proxy
llm-anygate-cli start

Step 4: Use the Proxy

Your proxy is now running at http://localhost:4567 with an OpenAI-compatible API:

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:4567/v1",
    api_key="sk-my-secure-key"
)

response = client.chat.completions.create(
    model="gpt-4o",  # or any model from your config
    messages=[{"role": "user", "content": "Hello!"}]
)

Generated Project Structure

The CLI generates a complete project with:

my-proxy/
├── config.yaml         # Full LiteLLM configuration
├── env.example         # Template for API keys
├── anygate.yaml       # Project configuration for llm-anygate-cli
├── README.md          # Project documentation
└── .gitignore         # Git ignore rules

CLI Usage

Create Command

llm-anygate-cli create [options]

Options:

• --project <dir> (required) - Directory to create the project in
• --model-config <file> (optional) - Path to your model configuration YAML (generates default gpt-4o config if not provided)
• --port <number> - Port for the proxy server (default: 4567)
• --master-key <key> - Master key for API authentication (default: sk-dummy)

Start Command

llm-anygate-cli start [options]

Options:

• --project <dir> (optional) - Project directory (default: current directory)
• --port <number> (optional) - Override port from project configuration
• --master-key <key> (optional) - Override master key from project configuration

The start command reads configuration from anygate.yaml in the project directory.

Examples

# Create with default configuration
llm-anygate-cli create --project my-proxy

# Create with custom configuration
llm-anygate-cli create \
  --project /path/to/my-llm-proxy \
  --model-config configs/models.yaml \
  --port 8080 \
  --master-key "sk-production-key-here"

# Start proxy from project directory
cd my-proxy
llm-anygate-cli start

# Start proxy with overrides
llm-anygate-cli start --port 3000 --master-key "sk-new-key"

Model Configuration Format

The model configuration is a simple YAML file with a model_list array:

model_list:
  - model_name: <name-for-your-app>
    litellm_params:
      model: <provider>/<model-id>
      api_base: <api-endpoint>  # Optional
      api_key: os.environ/<ENV_VAR_NAME>
      # Additional parameters as needed

Supported Providers

• OpenAI and OpenAI-compatible endpoints
• Anthropic (Claude)
• Google (Gemini/Vertex)
• Azure OpenAI
• Local models (Ollama, etc.)
• Any provider supported by LiteLLM

Why LLM AnyGate?

The Problem

Setting up LiteLLM proxy servers requires understanding complex configurations, database setups, and various deployment options. This complexity is a barrier for developers who just want a simple proxy for their AI tools.

The Solution

LLM AnyGate provides a simple CLI that generates everything you need:

• ✅ No database required (stateless operation)
• ✅ Minimal configuration needed
• ✅ Cross-platform start scripts
• ✅ Environment variable management
• ✅ Production-ready settings

Development

Project Structure

llm-anygate/
├── src/llm_anygate/       # Main package source code
│   ├── cli_tool.py        # CLI interface
│   ├── config_converter.py # Config conversion logic
│   ├── proxy_generator.py  # Project generation
│   └── templates.py        # File templates
├── tests/                  # Test suite
├── docs/                   # Documentation
└── context/                # AI collaboration workspace

Running Tests

pixi run test           # Run tests
pixi run test-cov       # Run tests with coverage

Code Quality

pixi run lint           # Run linting
pixi run format         # Format code
pixi run typecheck      # Type checking
pixi run quality        # Run all checks

Roadmap

• Core CLI tool implementation
• LiteLLM configuration generation
• Cross-platform CLI commands (create & start)
• Environment variable management
• PyPI package publishing
• Default configuration support
• Docker composition generator
• Provider connectivity testing
• Configuration validation
• Web UI for configuration
• Metrics and monitoring integration
• Advanced routing and load balancing

Requirements

• Python 3.11 or higher
• LiteLLM CLI tool (for running generated proxies)

  # Recommended: Install using uv
  uv tool install 'litellm[proxy]'
  
  # Alternative: Install with pip
  pip install 'litellm[proxy]'
  

Security Notes

• Generated projects include env.example as a template for API keys
• Never commit .env files with actual API keys
• Always use secure master keys in production
• Generated .gitignore excludes sensitive files

Contributing

Contributions are welcome! Please see our Contributing Guide for details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Built with OmegaConf for robust configuration handling
• Uses Pixi for environment management
• Generates configurations for LiteLLM
• Project structure based on magic-context templates

Contact

• GitHub: @igamenovoer
• Issues: GitHub Issues

Support

For questions, issues, or feature requests, please open an issue on GitHub.