Google AI Studio SDK

Pass-through endpoints for Google AI Studio - call provider-specific endpoint, in native format (no translation).

Just replace https://generativelanguage.googleapis.com with LITELLM_PROXY_BASE_URL/gemini

Example Usage

curl 'http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:countTokens?key=sk-anything' \\
-H 'Content-Type: application/json' \\
-d '{
    "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }]
        }]
}'

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

const ai = new GoogleGenAI({
    apiKey: "sk-1234", // litellm proxy API key
    httpOptions: {
        baseUrl: "http://localhost:4000/gemini", // http://<proxy-base-url>/gemini
    },
});

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

// For streaming responses
async function main_streaming() {
    try {
        const response = await ai.models.generateContentStream({
            model: "gemini-2.5-flash",
            contents: "Explain how AI works",
        });
        for await (const chunk of response) {
            process.stdout.write(chunk.text);
        }
    } catch (error) {
        console.error('Error:', error);
    }
}

main();
// main_streaming();

Supports ALL Google AI Studio Endpoints (including streaming).

See All Google AI Studio Endpoints

Quick Start

Let's call the Gemini /countTokens endpoint

1. Add Gemini API Key to your environment

2. Start LiteLLM Proxy

litellm

# RUNNING on http://0.0.0.0:4000

3. Test it!

Let's call the Google AI Studio token counting endpoint

http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:countTokens?key=anything' \\
-H 'Content-Type: application/json' \\
-d '{
    "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }]
        }]
}'

Examples

Anything after http://0.0.0.0:4000/gemini is treated as a provider-specific route, and handled accordingly.

Key Changes:

Example 1: Counting tokens

LiteLLM Proxy Call

curl http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:countTokens?key=anything \\
    -H 'Content-Type: application/json' \\
    -X POST \\
    -d '{
      "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }],
        }],
      }'

Direct Google AI Studio Call

curl https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:countTokens?key=$GOOGLE_API_KEY \\
    -H 'Content-Type: application/json' \\
    -X POST \\
    -d '{
      "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }],
        }],
      }'

Example 2: Generate content

LiteLLM Proxy Call

curl "http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:generateContent?key=anything" \\
    -H 'Content-Type: application/json' \\
    -X POST \\
    -d '{
      "contents": [{
        "parts":[{"text": "Write a story about a magic backpack."}]
        }]
       }' 2> /dev/null

Direct Google AI Studio Call

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent?key=$GOOGLE_API_KEY" \\
    -H 'Content-Type: application/json' \\
    -X POST \\
    -d '{
      "contents": [{
        "parts":[{"text": "Write a story about a magic backpack."}]
        }]
       }' 2> /dev/null

Example 3: Caching

curl -X POST "http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash-001:generateContent?key=anything" \\
-H 'Content-Type: application/json' \\
-d '{
      "contents": [
        {
          "parts":[{
            "text": "Please summarize this transcript"
          }],
          "role": "user"
        },
      ],
      "cachedContent": "'$CACHE_NAME'"
    }'

Direct Google AI Studio Call

curl -X POST "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash-001:generateContent?key=$GOOGLE_API_KEY" \\
-H 'Content-Type: application/json' \\
-d '{
      "contents": [
        {
          "parts":[{
            "text": "Please summarize this transcript"
          }],
          "role": "user"
        },
      ],
      "cachedContent": "'$CACHE_NAME'"
    }'

Example 4: Video Generation with Veo

Generate videos using Google's Veo model through LiteLLM pass-through routes.

→ Complete Veo Video Generation Guide

Advanced

Pre-requisites

• Setup proxy with DB

Use this, to avoid giving developers the raw Google AI Studio key, but still letting them use Google AI Studio endpoints.

Use with Virtual Keys

1. Setup environment

litellm

# RUNNING on http://0.0.0.0:4000

2. Generate virtual key

curl -X POST 'http://0.0.0.0:4000/key/generate' \\
-H 'Authorization: Bearer sk-1234' \\
-H 'Content-Type: application/json' \\
-d '{}'

Expected Response

{
    ...
    "key": "sk-1234ewknldferwedojwojw"
}

3. Test it!

http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:countTokens?key=sk-1234ewknldferwedojwojw' \\
-H 'Content-Type: application/json' \\
-d '{
    "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }]
        }]
}'

Send tags in request headers

Use this if you want tags to be tracked in the LiteLLM DB and on logging callbacks.

Pass tags in request headers as a comma separated list. In the example below the following tags will be tracked

tags: ["gemini-js-sdk", "pass-through-endpoint"]

curl 'http://0.0.0.0:4000/gemini/v1beta/models/gemini-1.5-flash:generateContent?key=sk-anything' \\
-H 'Content-Type: application/json' \\
-H 'tags: gemini-js-sdk,pass-through-endpoint' \\
-d '{
    "contents": [{
        "parts":[{
          "text": "The quick brown fox jumps over the lazy dog."
          }]
        }]
}'

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

const ai = new GoogleGenAI({
    apiKey: "sk-1234",
    httpOptions: {
        baseUrl: "http://localhost:4000/gemini", // http://<proxy-base-url>/gemini
        headers: {
            "tags": "gemini-js-sdk,pass-through-endpoint",
        },
    },
});

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

main();