Skip to content

Connect to Gemini from Python

Connect to Gemini from Python

This guide shows the clean path for calling the Gemini API from Python:

install the recommended SDK
store the API key securely
send a first request
keep the code easy to evolve into a real application

1. Install the SDK

python -m pip install google-genai python-dotenv

2. Store the API Key

Set the key as an environment variable.

Windows PowerShell

$env:GEMINI_API_KEY="your-api-key-here"

macOS or Linux

export GEMINI_API_KEY="your-api-key-here"

For local development, a .env file is also fine:

GEMINI_API_KEY=your-api-key-here

3. Make a First Request

from google import genai

client = genai.Client()

response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="Explain the difference between fine-tuning and RAG."
)

print(response.text)

Why This Example Uses `google-genai`

Google’s current Gemini API documentation recommends the Google GenAI SDK. Older tutorials may still use earlier package names and older API styles. This page intentionally uses the newer recommended client shape.

4. Add `.env` Support for Local Development

import os
from dotenv import load_dotenv
from google import genai

load_dotenv()

client = genai.Client(api_key=os.getenv("GEMINI_API_KEY"))

Common Mistakes

storing the key directly in source files
using an outdated SDK example without checking the current recommended package
assuming the same model name will remain the best choice for every workload

Practical Advice

keep authentication separate from business logic
treat model names as config, not hardcoded design decisions
start with a single request script before wiring Gemini into a larger app