Google GenAI SDK - ModRouter

Endpoint

https://api.modrouter.online/genai

Python

import os
from google import genai
from google.genai.types import HttpOptions

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

client = genai.Client(
    api_key=api_key,
    http_options=HttpOptions(base_url=f"{base_url}/genai"),
)

response = client.models.generate_content(
    model="gemini/gemini-2.5-flash",
    contents="Hello",
)

print(response.text)

JavaScript

Use the base URL option supported by your installed GenAI SDK version:

import { GoogleGenAI } from "@google/genai";

const baseURL = process.env.MODROUTER_BASE_URL.replace(/\/$/, "");
const apiKey = process.env.MODROUTER_API_KEY;

const ai = new GoogleGenAI({
  apiKey,
  httpOptions: {
    baseUrl: `${baseURL}/genai`,
  },
});

const response = await ai.models.generateContent({
  model: "gemini/gemini-2.5-flash",
  contents: "Hello",
});

console.log(response.text);

Supported Route Groups

ModRouter’s GenAI integration includes:

model generate and stream routes under /genai/v1beta/models/{model:*}
model retrieve and list routes
file upload, list, retrieve, delete
batch create, retrieve, list, cancel
cached content create, list, retrieve, update, delete
operation retrieve routes
rank route at /genai/v1/rank

This endpoint expects GenAI request shapes. Use /openai for OpenAI SDK request shapes and /anthropic for Anthropic SDK request shapes.

Models

Native Gemini names and explicit ModRouter model IDs both work where the route can resolve them:

gemini-2.5-flash
gemini/gemini-2.5-pro
vertex/gemini-2.5-pro
openai/gpt-4o-mini
anthropic/claude-sonnet-4-5

Non-Google providers must support the features required by the GenAI request you send.