lm-deluge

Unified async Python client for LLM APIs. Supports OpenAI, Anthropic, Google, Together, Mistral, Groq, and more.

Quick Reference

Imports

from lm_deluge import LLMClient, Conversation, Tool, APIResponse

Simple Request

llm = LLMClient(model_names="claude-4.5-haiku", max_new_tokens=1024)
response = await llm.start(Conversation().user("Hello!"))
print(response.completion)  # Text output - NOT .text