LiteLLM
Route LiteLLM completion calls through the gateway with one wrapper.
LiteLLM is an inference SDK. DeepintShield accepts its completion calls through an OpenAI-compatible route and applies normal inference policies to them.
What works
completion()andacompletion(), including streaming- Tools, timeouts and
extra_headers - The
shield.litellm()wrapper withdeepintshield[litellm]==3.2.3
Limits to know: The wrapper does not add Responses, embeddings, image, audio, file or video methods. LiteLLM calls do not create an Agentic tool registration.
How to connect
In native LiteLLM, call completion(model="openai/<provider>/<model>", custom_llm_provider="openai", api_base="https://app.deepintshield.com/litellm", api_key=...) with your Virtual Key; with the wrapper, drop the outer openai/ prefix.
Read the LiteLLM integration docs. Comparing the two? See DeepintShield vs LiteLLM.