Google Gen AI Python SDK: A Full Information

Generative AI fashions are altering how we create content material, whether or not it’s textual content, photographs, video, or code. With Google’s Gen AI Python SDK, now you can entry and work together with Google’s generative AI fashions in your Python purposes extra simply, along with utilizing the Gemini Developer API and Vertex AI APIs. Meaning builders can extra readily create purposes, together with chatbots, content material turbines, or artistic instruments. On this article, we are going to cowl all the pieces you want to know to get began utilizing the Google Gen AI Python SDK.

Additionally learn: Construct an LLM Mannequin utilizing Google Gemini API

What’s the Google Gen AI Python SDK?

The Google Gen AI Python SDK is a consumer library for builders to make use of Google’s generative AI skills simply utilizing Python. It gives:

• Help for Gemini Developer API (Google’s superior textual content and multimodal generative fashions)
• Integration with Vertex AI APIs for enterprise-scale AI workloads
• Help for producing textual content, photographs, movies, embeddings, chat conversations, and extra
• Instruments for file administration, caching, and async assist
• Superior operate calling and schema enforcement options

This SDK additionally abstracts a lot of the complexity round API calls and permits you to deal with constructing AI-powered purposes.

Set up

Putting in the SDK is straightforward. Run:

pip set up google-genai

The above command will set up the Google Gen AI Python SDK bundle utilizing pip. This command downloads all the pieces you want for the Python surroundings to start out up the Google generative AI providers, together with the sources and all dependencies.

Imports and Consumer Setup

After getting put in the SDK, create a Python file and import the SDK:

from google import genai
from google.genai import sorts

The SDK has two modules – genai and kinds. The genai module creates a consumer used for API interplay, whereas the kinds module has knowledge buildings and lessons that function helpers used to construct requests and configure request parameters.

You’ll create an occasion of the consumer for every interplay with the Google generative AI fashions. You’ll instantiate the consumer with completely different strategies relying on the API you’re utilizing.

For the Gemini Developer API, you’ll be able to instantiate the consumer by passing alongside your API key:

consumer = genai.Consumer(api_key='YOUR_GEMINI_API_KEY')

You instantiate the consumer you’ll be able to work together with the Gemini Developer API by passing in your API key. This consumer will maintain the entry token and request administration.

Non-obligatory: Utilizing Google Cloud Vertex AI

consumer = genai.Consumer(
   vertexai=True,
   venture="your-project-id",
   location='us-central1'
)

If you’re going to use Google Cloud Vertex AI, you’ll initialise the consumer in another way by specifying the venture ID and the situation.

Word: Utilizing Vertex AI is non-compulsory. You may create your venture ID right here.

If you don’t use Vertex AI, you’ll be able to merely use the API key methodology above.

API Model and Configuration

By default, the SDK makes use of beta endpoints to entry beta options. Nevertheless, if you wish to use steady APIs, you’ll be able to specify the API model utilizing the http_options argument:

from google.genai import sorts

consumer = genai.Consumer(
   vertexai=True,
   venture="your-project-id",
   location='us-central1',
   http_options=sorts.HttpOptions(api_version='v1')
)

It’s as much as you ways you wish to proceed to steadiness stability with cutting-edge options.

Utilizing Atmosphere Variables (Non-obligatory)

As a substitute of straight passing keys, we should always first set surroundings variables:

Gemini Developer API:

export GEMINI_API_KEY='your-api-key'

Vertex AI:

export GOOGLE_GENAI_USE_VERTEXAI=true
export GOOGLE_CLOUD_PROJECT='your-project-id'
export GOOGLE_CLOUD_LOCATION='us-central1'

Then, initialize the consumer merely with:

consumer = genai.Consumer()

Google Gen AI Python SDK Use Circumstances

Listed here are the assorted methods you’ll be able to put Google Gen AI Python SDK’s capabilities to make use of as soon as arrange.

Content material Era

The first operate of the SDK is to generate AI content material. You present prompts in varied varieties, equivalent to easy strings, structured content material, or complicated multimodal inputs.

Primary Textual content Era

response = consumer.fashions.generate_content(
   mannequin="gemini-2.0-flash-001",
   contents="Why Does the solar rises from east"
)
print(response.textual content)

Output

This sends a immediate to the mannequin and returns the generated reply.

Structured Content material Inputs

You may insert structured content material throughout varied roles, like consumer or mannequin for chatbot, conversational, or multi-turn contexts.

from google.genai import sorts


content material = sorts.Content material(
   position="consumer",
   elements=[types.Part.from_text(text="Tell me a fun fact about work.")]
)
response = consumer.fashions.generate_content(mannequin="gemini-2.0-flash-001", contents=content material)
print(response.textual content)

Output

The SDK internally interprets many various enter sorts to a structured knowledge format for the mannequin.

File Add and Utilization

The Gemini Builders API permits you to add information for the mannequin to course of. That is nice for summarization, or content material extraction:

file = consumer.information.add(file="/content material/sample_file.txt")
response = consumer.fashions.generate_content(
   mannequin="gemini-2.0-flash-001",
   contents=[file, 'Please summarize this file.']
)
print(response.textual content)

Output

This is a perfect strategy for including AI performance to document-based duties.

Operate Calling

A singular functionality is the power to move Python capabilities as “instruments” for the mannequin to invoke robotically whereas producing the completion.

def get_current_weather(location: str) -> str:
   return 'sunny'


response = consumer.fashions.generate_content(
   mannequin="gemini-2.0-flash-001",
   contents="What's the climate like in Ranchi?",
   config=sorts.GenerateContentConfig(instruments=[get_current_weather])
)
print(response.textual content)

Output

This permits dynamic, real-time knowledge integration inside AI responses.

Superior Configuration

You’ve the power to customise technology with parameters equivalent to temperature, max_output_tokens, and security settings to handle randomness, size, and filter dangerous content material.

config = sorts.GenerateContentConfig(
   temperature=0.3,
   max_output_tokens=100,
   safety_settings=[types.SafetySetting(category='HARM_CATEGORY_HATE_SPEECH', threshold='BLOCK_ONLY_HIGH')]
)


response = consumer.fashions.generate_content(
   mannequin="gemini-2.0-flash-001",
   contents=""'Provide some encouraging phrases for somebody beginning a brand new journey.''',
   config=config
)
print(response.textual content)

Output

This may present granularity over content material high quality and security.

Multimedia Help: Photos and Movies

The SDK permits you to generate and edit photographs and generate movies (in preview).

• Generate photographs utilizing textual content prompts.
• Upscale or modify photographs generated.
• Generate movies from textual content or photographs.

Instance of Picture Era:

response = consumer.fashions.generate_images(
   mannequin="imagen-3.0-generate-002",
   immediate="A tranquil seashore with crystal-clear water and colourful seashells on the shore.",
   config=sorts.GenerateImagesConfig(number_of_images=1)
)
response.generated_images[0].picture.present()

Output

Instance of Video Era:

import time
operation = consumer.fashions.generate_videos(
   mannequin="veo-2.0-generate-001",
   immediate="A cat DJ spinning vinyl data at a futuristic nightclub with holographic beats.",
   config=sorts.GenerateVideosConfig(number_of_videos=1, duration_seconds=5)
)


whereas not operation.achieved:
   time.sleep(20)
   operation = consumer.operations.get(operation)
video = operation.response.generated_videos[0].video
video.present()

Output:

This permits for artistic, multimodal AI apps.

Chat and Conversations

You can begin chat classes that protect context all through your messages:

chat = consumer.chats.create(mannequin="gemini-2.0-flash-001")
response = chat.send_message('Inform me a narrative')
print(response.textual content)

response = chat.send_message('Summarize that story in a single sentence')
print(response.textual content)

That is helpful for creating conversational AI that remembers earlier dialogue.

Asynchronous Help

All predominant API strategies have async capabilities for higher integration into async Python apps:

response = await consumer.aio.fashions.generate_content(
   mannequin="gemini-2.0-flash-001",
   contents="Inform a Horror story in 200 phrases."
)
print(response.textual content)

Token Counting

Token counting tracks what number of tokens (items of textual content) are in your enter. This helps you keep inside mannequin limits and make cost-effective selections.

token_count = consumer.fashions.count_tokens(
   mannequin="gemini-2.0-flash-001",
   contents="Why does the sky have a blue hue as an alternative of different colours?"
)
print(token_count)

Embeddings

Embeddings flip your textual content into numeric vectors that characterize its that means, which can be utilized for search, clustering, and AI analysis.

embedding = consumer.fashions.embed_content(
   mannequin="text-embedding-004",
   contents="Why does the sky have a blue hue as an alternative of different colours?"
)
print(embedding)

Utilizing the SDK, you’ll be able to simply depend tokens and make embeddings to enhance and improve your AI purposes.

Conclusion

The Google Gen AI Python SDK is a sturdy, versatile instrument that enables builders to entry Google’s high generative AI fashions of their Python initiatives. From textual content technology, chat, and chatbot, to picture/video technology, operate calling, and way more it gives a sturdy function set with easy interfaces. With a straightforward bundle set up, easy consumer configuration course of, and assist for async programming and multimedia, the SDK makes constructing purposes that leverage AI considerably simpler. Whether or not you’re a newbie or seasoned developer, utilizing the SDK is comparatively painless however highly effective in terms of incorporating generative AI into your workflows.

Ceaselessly Requested Questions

Hello, I’m Janvi, a passionate knowledge science fanatic presently working at Analytics Vidhya. My journey into the world of knowledge started with a deep curiosity about how we are able to extract significant insights from complicated datasets.