langchain-google-vertexai

This package contains the LangChain integrations for Google Cloud generative models.

Contents

• langchain-google-vertexai
  • Contents
  • Installation
  • Chat Models
    • Multimodal inputs
    • Multimodal Outputs
  • Embeddings
  • LLMs
  • Code Generation
    • Example: Generate a Python function
    • Example: Generate JavaScript code
    • Notes

Installation

pip install -U langchain-google-vertexai

Chat Models

ChatVertexAI class exposes models such as gemini-pro and other Gemini variants.

To use, you should have a Google Cloud project with APIs enabled, and configured credentials. Initialize the model as:

from langchain_google_vertexai import ChatVertexAI

llm = ChatVertexAI(model_name="gemini-2.5-flash")
llm.invoke("Sing a ballad of LangChain.")

Multimodal inputs

Gemini supports image inputs when providing a single chat message. Example:

from langchain_core.messages import HumanMessage
from langchain_google_vertexai import ChatVertexAI

llm = ChatVertexAI(model_name="gemini-2.0-flash-001")
message = HumanMessage(
    content=[
        {
            "type": "text",
            "text": "What's in this image?",
        },
        {
            "type": "image_url",
            "image_url": {"url": "https://picsum.photos/seed/picsum/200/300"}
        },
    ]
)
response = llm.invoke([message])

The value of image_url can be:

• A public image URL
• An accessible Google Cloud Storage (GCS) file (e.g., "gcs://path/to/file.png")
• A base64 encoded image (e.g., "data:image/png;base64,abcd124")

Multimodal Outputs

Gemini supports image output. Example:

from langchain_core.messages import HumanMessage
from langchain_google_vertexai import ChatVertexAI, Modality

llm = ChatVertexAI(model_name="imagen-3.0-generate-002",
                   response_modalities = [Modality.TEXT, Modality.IMAGE])
message = HumanMessage(
    content=[
        {
            "type": "text",
            "text": "Generate an image of a cat.",
        },
    ]
)
response = llm.invoke([message])

Embeddings

Google Cloud embeddings models can be used as:

from langchain_google_vertexai import VertexAIEmbeddings

embeddings = VertexAIEmbeddings()
embeddings.embed_query("hello, world!")

LLMs

Use Google Cloud's generative AI models as old-style LangChain LLMs:

from langchain_core.prompts import PromptTemplate
from langchain_google_vertexai import ChatVertexAI

template = """Question: {question}

Answer: Let's think step by step."""
prompt = PromptTemplate.from_template(template)

llm = ChatVertexAI(model_name="gemini-2.5-flash")
chain = prompt | llm

question = "Who was the president of the USA in 1994?"
print(chain.invoke({"question": question}))

Code Generation

You can use Gemini models for code generation tasks to generate code snippets, functions, or scripts in various programming languages.

Example: Generate a Python function

from langchain_google_vertexai import ChatVertexAI

llm = ChatVertexAI(model_name="gemini-2.5-flash", temperature=0.3, max_output_tokens=1000)

prompt = "Write a Python function that checks if a string is a valid email address."

generated_code = llm.invoke(prompt)
print(generated_code)

Example: Generate JavaScript code

from langchain_google_vertexai import ChatVertexAI

llm = ChatVertexAI(model_name="gemini-2.5-flash", temperature=0.3, max_output_tokens=1000)
prompt_js = "Write a JavaScript function that returns the factorial of a number."

print(llm.invoke(prompt_js))

Notes

• Adjust temperature to control creativity (higher values increase randomness).
• Use max_output_tokens to limit the length of the generated code.
• Gemini models are well-suited for code generation tasks with advanced understanding of programming concepts.