Providers

autochecklist supports multiple LLM providers through one common API.

Which Provider Should You Use?

Provider	Best for	Required env var
openrouter (default)	Access many models through one key	OPENROUTER_API_KEY
openai	Direct OpenAI usage	OPENAI_API_KEY
vllm	Self-hosted OpenAI-compatible endpoint	Usually none (depends on your server)

OpenRouter (Default)

from autochecklist import pipeline

pipe = pipeline("tick", generator_model="openai/gpt-5-mini", scorer_model="openai/gpt-5-mini")
# Uses OPENROUTER_API_KEY from environment

OpenAI Direct

from autochecklist import pipeline

pipe = pipeline("tick", provider="openai", generator_model="gpt-5-mini")
# Uses OPENAI_API_KEY from environment

vLLM Server Mode

Connect to any OpenAI-compatible endpoint:

from autochecklist import pipeline

pipe = pipeline("tick", provider="vllm", base_url="http://gpu:8000/v1", generator_model="meta-llama/Llama-3-8B")

vLLM Offline (No Server)

Direct Python inference without a server:

from autochecklist import VLLMOfflineClient, pipeline

client = VLLMOfflineClient(model="google/gemma-3-1b-it")
pipe = pipeline("tick", client=client)

API Format: Chat vs Responses

Opt-in to OpenAI's Responses API format:

from autochecklist import pipeline

pipe = pipeline("tick", provider="openai", generator_model="gpt-5-mini", api_format="responses")

Default is api_format="chat". Keep that unless you specifically need Responses API behavior.