Google Gemini API

Pro v1.0.0

Access Google's multimodal AI for text generation, vision, image generation, function calling, and advanced reasoning with Gemini models

Requires Claude Code CLI

This skill integrates with Google Gemini through Claude Code. Install Claude Code and add this skill to use it.

Available Actions

Interactive Chat (REPL)

Start an interactive conversation with Gemini in your terminal with access to built-in tools, file uploads, and MCP servers

Try saying...

  • "Start a conversation with Gemini"
  • "Chat with Gemini interactively"
  • "Launch the Gemini REPL"

Generate Text Content

Generate text responses from a single prompt without entering interactive mode

Parameters

Name Type Required Description
prompt string Required The text prompt to send to Gemini
model string Optional Specific model to use

Try saying...

  • "Generate text about quantum computing"
  • "Ask Gemini to write a haiku about programming"
  • "Get a quick explanation of machine learning"

Analyze Images (Vision)

Use Gemini's multimodal vision capabilities to understand, describe, and extract information from images

Parameters

Name Type Required Description
file string Required Path to image file (JPG, PNG, WebP, etc.)
prompt string Required Question or instruction about the image
media-resolution string Optional Resolution quality for image processing (Gemini 3+)

Try saying...

  • "Analyze this screenshot and extract all text"
  • "What's in this image?"
  • "Describe this technical diagram"
  • "Extract data from this chart image"

Generate Images

Create and edit images using Gemini's Nano Banana models with conversational iteration

Parameters

Name Type Required Description
prompt string Required Description of the image to generate
model string Optional Image generation model to use

Try saying...

  • "Generate an image of a futuristic city"
  • "Create an illustration of a dragon"
  • "Make an infographic about renewable energy"
  • "Generate an image with text 'Welcome to AI'"

Process Documents

Analyze PDFs and documents (up to 1000 pages) with full text extraction and understanding

Parameters

Name Type Required Description
file string Required Path to PDF or document file
prompt string Required Question or task about the document

Try saying...

  • "Analyze this PDF and summarize it"
  • "Extract key points from this research paper"
  • "What are the main conclusions in this report?"
  • "List all obligations in this contract"

Function Calling

Enable Gemini to call external tools and APIs defined in your configuration

Parameters

Name Type Required Description
tools object Required Tool definitions in settings.json with name, description, and parameters

Try saying...

  • "Get the current weather"
  • "Search for information using my custom tool"
  • "Call the database query function"

Google Search Grounding

Ground Gemini's responses in real-time Google Search results with citations

Parameters

Name Type Required Description
query string Required Search query or question requiring current information

Try saying...

  • "Search for the latest AI news"
  • "What are recent developments in quantum computing?"
  • "Find current statistics on climate change"

Code Execution

Let Gemini write and execute code to solve problems, with iterative debugging

Parameters

Name Type Required Description
task string Required Programming task or problem to solve

Try saying...

  • "Write Python code to analyze this data"
  • "Calculate the fibonacci sequence"
  • "Debug this code snippet"
  • "Write a function to parse JSON"

Long Context Processing

Process large codebases, documents, or multiple files using Gemini's 1M+ token context window

Parameters

Name Type Required Description
file string Required Large file or archive to process
prompt string Required Analysis task for the large context

Try saying...

  • "Analyze this entire codebase"
  • "Review these 500 pages of documentation"
  • "Process all these files together"

Getting Started

API Key Easy Setup
  1. 1 Go to Google AI Studio: https://aistudio.google.com/apikey
  2. 2 Click 'Create API Key'
  3. 3 Choose to create in new or existing Google Cloud project
  4. 4 Copy the API key (shown only once)
  5. 5 Add to shell config: echo 'export GEMINI_API_KEY="YOUR_KEY"' >> ~/.zshrc
  6. 6 Restart terminal or run: source ~/.zshrc

Verify Setup

gemini --version && [ -n "$GEMINI_API_KEY" ] && echo "API key configured" || echo "API key not found"

Success: CLI version number displayed and 'API key configured' message

View full setup guide

Security & Access

Access Scope

Full access to Gemini API under your Google Cloud project. All API calls are billed to your account. Can generate content, process uploaded files, call configured functions, and access search results.

  • Never commit GEMINI_API_KEY to version control
  • API key grants full billing access to your Google Cloud project
  • Store keys in environment variables, not in code files
  • Monitor usage in Google AI Studio to detect unauthorized access
  • Use separate API keys for development and production environments
  • Content sent to the API may be used to improve Google's services
  • Generated images include SynthID watermarks for tracking
  • Function calling can execute arbitrary code if misconfigured
  • Rate limit exceeded errors may indicate key compromise

Limitations

  • Rate limits vary by account tier (free vs pay-as-you-go)
  • Image generation includes SynthID watermark
  • Some features require Gemini 3 models
  • Files have size limits (use Files API for very large files)
  • Google Search grounding is billed separately (since Jan 5, 2026)
  • Content may be used to improve Google services (check ToS)

Get This Skill

Requires Pro subscription ($9/month)

Quick Reference

Type
CLI
Auth
API Key
Setup
Easy Setup
Tools Required
Bash
CLI Dependencies
gemini node npm

Documentation

Have Feedback?

Help us improve this skill by sharing your ideas and suggestions.

Request Improvements