llm-clients

A reusable library for LLM clients with LangChain.

Supports OpenAI, Anthropic, Azure OpenAI, Google Gemini, and HuggingFace.

Install

# Core only
pip install llm-clients

# With specific provider(s)
pip install "llm-clients[openai]"
pip install "llm-clients[anthropic]"
pip install "llm-clients[azure]"
pip install "llm-clients[google]"
pip install "llm-clients[huggingface]"

# All providers
pip install "llm-clients[all]"

Quick start

from llm_clients import LLMClientFactory

# Create any LLM (reads API keys from env vars by default)
llm = LLMClientFactory.create("openai", model="gpt-4o")
llm = LLMClientFactory.create("anthropic")
llm = LLMClientFactory.create("azure", azure_deployment="...", azure_endpoint="...")
llm = LLMClientFactory.create("google", model="gemini-2.0-flash")
llm = LLMClientFactory.create("huggingface", repo_id="HuggingFaceH4/zephyr-7b-beta")

Environment variables