Google Gemini

Google Gemini integration allows your product to connect with Google's multimodal AI models (Gemini 1.5 Pro, 1.5 Flash, and future versions) via the Google AI API (formerly PaLM / Bard API). This integration enables your product to perform text generation, summarization, code reasoning, image and file analysis, embeddings, and chat-based applications using Google's state-of-the-art Gemini models.

Credentials Needed

To integrate with the Google Gemini API, you need an API Key from Google AI Studio or a Service Account if you are integrating via Google Cloud.

Option 1 — API Key (Simpler Method):

• Google AI API Key

Option 2 — Service Account (Enterprise Integration):

• Service Account JSON Key File containing:
  • project_id
  • client_email
  • private_key

Use the API Key method for lightweight integrations, and Service Account method for secured, multi-user, or enterprise scenarios.

Permissions Needed / API Scopes

Gemini APIs use Google Cloud IAM and OAuth scopes for access control.

Functionality	Scope / Permission	Description
Generate text / chat	https://www.googleapis.com/auth/generative-language	Allows text and chat generation
Generate multimodal content (text + image)	Same as above	Used for Gemini 1.5 Pro and multimodal input
Embeddings	https://www.googleapis.com/auth/generative-language.retrieval	Allows embedding and retrieval APIs
API Key (no OAuth)	Direct API key auth	Simplified key-based authentication for most use cases

For most integrations, you only need API Key access — OAuth scopes apply mainly for user-based or enterprise flows.

Creating Users / Access Tokens

Step 1: Generate an API Key (Recommended for Most Use Cases)

1. Visit Google AI Studio: https://aistudio.google.com/
2. Sign in with your Google account.
3. Go to Get API key under the left panel or visit: https://aistudio.google.com/app/apikey
4. Click Create API key.
5. Copy the generated API key — it will start with AIza....

Step 2 (Alternative): Create a Service Account (Enterprise)

1. Go to Google Cloud Console → IAM & Admin → Service Accounts.
2. Click Create Service Account.
3. Provide a name (e.g., GeminiIntegrationSA).
4. Assign the following role:
   • Vertex AI User (roles/aiplatform.user)
5. Click Continue → Done.
6. Open the created service account → Keys → Add Key → Create New Key → JSON.
7. Download the key file securely (e.g., gemini-service-key.json).

Test Connectivity

You can test Gemini using curl, the Google Generative AI SDK, or directly through the REST API:

Example: Using curl

curl \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: <GEMINI_API_KEY>" \
  https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent \
  -d '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "Hello Gemini! Can you confirm connectivity?"}]
    }]
  }'

If successful, you'll receive a JSON response with model output.

Example: Using Python SDK

from google import genai

client = genai.Client(api_key="<GEMINI_API_KEY>")
response = client.models.generate_content(
    model="gemini-1.5-flash",
    contents="Hello Gemini, this is a test!"
)
print(response.text)

Save the Results in the Platform and Create Connection

In your platform's integration module, securely store:

• GEMINI_API_KEY (for API key auth)
• or SERVICE_ACCOUNT_JSON (for enterprise auth)

Label the connection as Google Gemini Integration.

Test the connection by generating sample text or running a small chat completion request.

Best Practices

• Use API keys for single-tenant or developer integrations; use Service Accounts for multi-tenant or enterprise setups.
• Store credentials in a secret vault or Google Secret Manager, never in plain text.
• Rotate API keys periodically.
• Choose the right model:
  • Gemini 1.5 Flash — fast and cost-effective for chat, summarization, or light reasoning.
  • Gemini 1.5 Pro — for deeper reasoning, multimodal (text, image, code) tasks.
• Handle rate limits gracefully — check error codes 429 (Too Many Requests).
• Log token usage and monitor billing in the Google Cloud Console → Billing → Reports.
• For multimodal input (image, text, file), ensure proper encoding (Base64 for images).
• Follow Google AI safety and usage policies.