LiteLLM - ModRouter

Endpoint

https://api.modrouter.online/litellm

ModRouter’s LiteLLM router combines OpenAI, Anthropic, GenAI, Bedrock, and Cohere route families under the /litellm prefix.

Python

import os
from litellm import completion

base_url = os.environ["MODROUTER_BASE_URL"].rstrip("/")
api_key = os.environ["MODROUTER_API_KEY"]

response = completion(
    model="openai/gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
    api_base=f"{base_url}/litellm",
    api_key=api_key,
)

print(response.choices[0].message.content)

Provider Compatibility

LiteLLM already abstracts provider request shapes. ModRouter adds gateway features such as provider key management, fallbacks, logging, governance, and MCP tool injection where applicable. Requests still need to match a route family ModRouter implements under /litellm. If a LiteLLM provider or endpoint is not covered by ModRouter’s integration routes, use a ModRouter native endpoint or passthrough route instead.

Headers

Pass ModRouter headers through LiteLLM when you need virtual keys or routing metadata:

import os
from litellm import completion

base_url = os.environ["MODROUTER_BASE_URL"].rstrip("/")

response = completion(
    model="anthropic/claude-sonnet-4-5",
    messages=[{"role": "user", "content": "Hello"}],
    api_base=f"{base_url}/litellm",
    api_key="placeholder",
    extra_headers={"x-bf-vk": os.environ["MODROUTER_API_KEY"]},
)

​Endpoint

​Python

​Provider Compatibility

​Headers

Endpoint

Python

Provider Compatibility

Headers