Image Generation: AI Art with Gemini Models

PocketPaw can generate images using Google’s Gemini models via the google-genai SDK.

Setup

curl -fsSL https://pocketpaw.xyz/install.sh | sh

# Or add the image extra manually
pip install pocketpaw[image]

export POCKETPAW_GOOGLE_API_KEY="your-google-api-key"

Configuration

Usage

User: Generate an image of a cute robot holding a flower
Agent: [uses image_gen tool] → [generated image]

The tool returns the generated image which can be displayed in the web dashboard or sent via messaging channels.

Tool Schema

{
  "name": "image_gen",
  "description": "Generate an image from a text description",
  "input_schema": {
    "type": "object",
    "properties": {
      "prompt": {
        "type": "string",
        "description": "Description of the image to generate"
      }
    },
    "required": ["prompt"]
  }
}

Policy Group

Belongs to group:media. Control access with:

export POCKETPAW_TOOLS_ALLOW="group:media"

Installation

Requires the image extra:

curl -fsSL https://pocketpaw.xyz/install.sh | sh

# Or add the image extra manually
pip install pocketpaw[image]

This installs google-genai as an optional dependency.

Related