zen-completions 0.1.0

A CLI tool for interacting with LLMs through a model router

Zen Completions

A command-line interface for interacting with various LLM providers through a unified model router.

Features

• Support for multiple LLM providers (OpenAI, Azure OpenAI)
• Interactive chat mode
• Single completion mode with system prompt support
• Consistent interface across different models

Installation

Using pip

pip install git+https://github.com/zenafide/zen-completions.git

Using poetry

poetry add git+https://github.com/zenafide/zen-completions.git

Environment Variables

Configure your API keys using environment variables:

# For OpenAI
export OPENAI_API_KEY=your_openai_api_key

# For Azure OpenAI
export AZURE_OPENAI_API_KEY=your_azure_api_key
export AZURE_OPENAI_API_VERSION=2024-02-15-preview  # Optional
export AZURE_OPENAI_API_BASE=your_base_url  # Optional
export AZURE_OPENAI_DEPLOYMENT_NAME=your_deployment  # Optional
export AZURE_OPENAI_MODEL_NAME=your_model_name  # Optional

Usage

Interactive Chat

Start an interactive chat session:

zen chat

With a system prompt:

zen chat system "You are a coding assistant specialized in Python"

Single Completion

Get a single completion:

zen complete "What is the capital of France?"

With a system prompt:

zen complete "What is the capital of France?" system "You are a geographical expert"

Models and Options

Both commands support these options:

• --model or -m: Select the model to use
• --temperature or -t: Set the temperature for generation
• --max-tokens or -mt: Set the maximum tokens for the response

Example:

zen complete "Write a Python function to calculate factorial" --model gpt-4o --temperature 0.7 --max-tokens 500

License

MIT