Google GenAI SDK with LiteLLM

Use Google's official GenAI SDK (JavaScript/TypeScript and Python) with any LLM provider through LiteLLM Proxy.

The Google GenAI SDK (@google/genai for JS, google-genai for Python) provides a native interface for calling Gemini models. By pointing it to LiteLLM, you can use the same SDK with OpenAI, Anthropic, Bedrock, Azure, Vertex AI, or any other provider — while keeping the native Gemini request/response format.

Why Use LiteLLM with Google GenAI SDK?

Developer Benefits:

• Universal Model Access: Use any LiteLLM-supported model (Anthropic, OpenAI, Vertex AI, Bedrock, etc.) through the Google GenAI SDK interface
• Higher Rate Limits & Reliability: Load balance across multiple models and providers to avoid hitting individual provider limits, with fallbacks to ensure you get responses even if one provider fails

Proxy Admin Benefits:

• Centralized Management: Control access to all models through a single LiteLLM proxy instance without giving developers API keys to each provider
• Budget Controls: Set spending limits and track costs across all SDK usage
• Logging & Observability: Track all requests with cost tracking, logging, and analytics

Quick Start

1. Install the SDK

npm install @google/genai

uv add google-genai

2. Start LiteLLM Proxy

model_list:
  - model_name: gemini-2.5-flash
    litellm_params:
      model: gemini/gemini-2.5-flash
      api_key: os.environ/GEMINI_API_KEY

litellm --config config.yaml

3. Call the SDK through LiteLLM

const { GoogleGenAI } = require("@google/genai");

const ai = new GoogleGenAI({
  apiKey: "sk-1234",  // LiteLLM virtual key (not a Google key)
  httpOptions: {
    baseUrl: "http://localhost:4000/gemini",  // LiteLLM proxy URL
  },
});

async function main() {
  const response = await ai.models.generateContent({
    model: "gemini-2.5-flash",
    contents: "Explain how AI works",
  });
  console.log(response.text);
}

main();

from google import genai

client = genai.Client(
    api_key="sk-1234",  # LiteLLM virtual key (not a Google key)
    http_options={"base_url": "http://localhost:4000/gemini"},  # LiteLLM proxy URL
)

response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="Explain how AI works",
)
print(response.text)

curl "http://localhost:4000/gemini/v1beta/models/gemini-2.5-flash:generateContent?key=sk-1234" \\
  -H 'Content-Type: application/json' \\
  -X POST \\
  -d '{
    "contents": [{
      "parts": [{"text": "Explain how AI works"}]
    }]
  }'

Streaming

const { GoogleGenAI } = require("@google/genai");

const ai = new GoogleGenAI({
  apiKey: "sk-1234",
  httpOptions: {
    baseUrl: "http://localhost:4000/gemini",
  },
});

async function main() {
  const response = await ai.models.generateContentStream({
    model: "gemini-2.5-flash",
    contents: "Write a short poem about the ocean",
  });

  for await (const chunk of response) {
    process.stdout.write(chunk.text);
  }
}

main();

from google import genai

client = genai.Client(
    api_key="sk-1234",
    http_options={"base_url": "http://localhost:4000/gemini"},
)

response = client.models.generate_content_stream(
    model="gemini-2.5-flash",
    contents="Write a short poem about the ocean",
)

for chunk in response:
    print(chunk.text, end="")

Multi-turn Chat

const { GoogleGenAI } = require("@google/genai");

const ai = new GoogleGenAI({
  apiKey: "sk-1234",
  httpOptions: {
    baseUrl: "http://localhost:4000/gemini",
  },
});

async function main() {
  const chat = ai.chats.create({
    model: "gemini-2.5-flash",
  });

  const response1 = await chat.sendMessage({ message: "I have 2 dogs and 3 cats." });
  console.log(response1.text);

  const response2 = await chat.sendMessage({ message: "How many pets is that in total?" });
  console.log(response2.text);
}

main();

from google import genai

client = genai.Client(
    api_key="sk-1234",
    http_options={"base_url": "http://localhost:4000/gemini"},
)

chat = client.chats.create(model="gemini-2.5-flash")

response1 = chat.send_message("I have 2 dogs and 3 cats.")
print(response1.text)

response2 = chat.send_message("How many pets is that in total?")
print(response2.text)

Advanced: Use Any Model with the GenAI SDK

By default, the GenAI SDK talks to Gemini models. But with LiteLLM's router, you can route GenAI SDK requests to any provider — Anthropic, OpenAI, Bedrock, etc.

This works by using model_group_alias to map Gemini model names to your desired provider models. LiteLLM handles the format translation internally.

:::info

For this to work, point the SDK baseUrl to http://localhost:4000 (without /gemini). This routes requests through LiteLLM's native Google endpoints, which go through the router and support model aliasing.

:::

Route gemini-2.5-flash requests to Claude Sonnet:

model_list:
  - model_name: claude-sonnet
    litellm_params:
      model: anthropic/claude-sonnet-4-20250514
      api_key: os.environ/ANTHROPIC_API_KEY

router_settings:
  model_group_alias: {"gemini-2.5-flash": "claude-sonnet"}

Route gemini-2.5-flash requests to GPT-4o:

model_list:
  - model_name: gpt-4o-model
    litellm_params:
      model: gpt-4o
      api_key: os.environ/OPENAI_API_KEY

router_settings:
  model_group_alias: {"gemini-2.5-flash": "gpt-4o-model"}

Route gemini-2.5-flash requests to Claude on Bedrock:

model_list:
  - model_name: bedrock-claude
    litellm_params:
      model: bedrock/anthropic.claude-haiku-4-5-20251001:0
      aws_access_key_id: os.environ/AWS_ACCESS_KEY_ID
      aws_secret_access_key: os.environ/AWS_SECRET_ACCESS_KEY
      aws_region_name: us-east-1

router_settings:
  model_group_alias: {"gemini-2.5-flash": "bedrock-claude"}

Load balance across Anthropic and OpenAI:

model_list:
  - model_name: my-model
    litellm_params:
      model: anthropic/claude-sonnet-4-20250514
      api_key: os.environ/ANTHROPIC_API_KEY
  - model_name: my-model
    litellm_params:
      model: gpt-4o
      api_key: os.environ/OPENAI_API_KEY

router_settings:
  model_group_alias: {"gemini-2.5-flash": "my-model"}

Then use the SDK with baseUrl pointing to LiteLLM (without /gemini):

const { GoogleGenAI } = require("@google/genai");

const ai = new GoogleGenAI({
  apiKey: "sk-1234",
  httpOptions: {
    baseUrl: "http://localhost:4000",  // No /gemini — goes through the router
  },
});

async function main() {
  // This calls Claude/GPT-4o/Bedrock under the hood via model_group_alias
  const response = await ai.models.generateContent({
    model: "gemini-2.5-flash",
    contents: "Hello from any model!",
  });
  console.log(response.text);
}

main();

from google import genai

client = genai.Client(
    api_key="sk-1234",
    http_options={"base_url": "http://localhost:4000"},  # No /gemini
)

# This calls Claude/GPT-4o/Bedrock under the hood via model_group_alias
response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="Hello from any model!",
)
print(response.text)

Pass-through vs Native Router Endpoints

LiteLLM offers two ways to handle GenAI SDK requests:

Environment Variable Configuration

You can also configure the SDK via environment variables instead of code:

# For JavaScript SDK (@google/genai)

# For Python SDK (google-genai)
# Note: The Python SDK does not support a base URL env var.
# Configure it in code with http_options={"base_url": "..."} instead.

This is especially useful for tools built on top of the GenAI SDK (like Gemini CLI).

Related Resources

• Gemini CLI with LiteLLM
• Google AI Studio Pass-Through
• Google ADK with LiteLLM
• LiteLLM Proxy Quick Start
• @google/genai npm package
• google-genai PyPI package