Google AI (Gemini + Imagen)

The Google plugin provides comprehensive AI capabilities using Google’s Gemini and Imagen models.

Features

Feature	Provider	Description
Text Generation	Gemini	Generate text with structured JSON output support
Vision Analysis	Gemini	Analyze images and extract information
Image Editing	Gemini (Nano Banana)	Edit images using AI with natural language
Image Generation	Gemini / Imagen	Generate images from text prompts

Installation

pnpm add @teamflojo/floimg-google

API Key

Get your API key from Google AI Studio.

export GOOGLE_AI_API_KEY=your-api-key

All providers support per-request API keys via the apiKey parameter, allowing users to provide their own keys.

---

Gemini Text

Generate text using Gemini with optional structured JSON output.

Registration

import createClient from '@teamflojo/floimg';
import { geminiText } from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerTextProvider(geminiText());

Usage

// Simple text generation
const result = await floimg.generateText({
  provider: 'gemini-text',
  params: {
    prompt: 'Write 3 creative image prompts for a fantasy landscape',
  }
});

console.log(result.content);

Structured JSON Output

// Generate structured JSON
const result = await floimg.generateText({
  provider: 'gemini-text',
  params: {
    prompt: 'Generate image prompts for a product photoshoot',
    outputFormat: 'json',
    jsonSchema: {
      type: 'object',
      properties: {
        prompts: {
          type: 'array',
          items: { type: 'string' }
        },
        style: { type: 'string' },
        mood: { type: 'string' }
      }
    }
  }
});

console.log(result.parsed); // Typed JSON object

Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Text prompt for generation
systemPrompt	string	No	System prompt to guide behavior
context	string	No	Additional context (e.g., from vision analysis)
outputFormat	'text' | 'json'	No	Output format (default: 'text')
jsonSchema	object	No	JSON schema for structured output
temperature	number	No	Creativity (0-2, default: 0.7)
maxTokens	number	No	Max response tokens (default: 1000)
apiKey	string	No	Per-request API key override

---

Gemini Vision

Analyze images using Gemini’s multimodal capabilities.

Registration

import createClient from '@teamflojo/floimg';
import { geminiVision } from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerVisionProvider(geminiVision());

Usage

// Analyze an image
const result = await floimg.analyzeImage({
  provider: 'gemini-vision',
  blob: imageBlob,
  params: {
    prompt: 'What objects are in this image? List them with confidence scores.',
  }
});

console.log(result.content);

Structured Output

// Get structured analysis
const result = await floimg.analyzeImage({
  provider: 'gemini-vision',
  blob: imageBlob,
  params: {
    prompt: 'Analyze this product image',
    outputFormat: 'json',
    jsonSchema: {
      type: 'object',
      properties: {
        objects: { type: 'array', items: { type: 'string' } },
        colors: { type: 'array', items: { type: 'string' } },
        mood: { type: 'string' },
        suggestedTags: { type: 'array', items: { type: 'string' } }
      }
    }
  }
});

console.log(result.parsed.suggestedTags);

Parameters

Parameter	Type	Required	Description
prompt	string	No	What to analyze (default: “Describe this image”)
outputFormat	'text' | 'json'	No	Output format (default: 'text')
jsonSchema	object	No	JSON schema for structured output
maxTokens	number	No	Max response tokens (default: 1000)
apiKey	string	No	Per-request API key override

---

Gemini Edit (Nano Banana)

Edit images using Gemini’s native image generation with natural language instructions.

Nano Banana is Google’s codename for Gemini’s image generation/editing capability.

Registration

import createClient from '@teamflojo/floimg';
import { geminiTransform } from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerTransformProvider(geminiTransform());

Usage

// Edit an image
const edited = await floimg.transform({
  blob: inputImage,
  op: 'edit',
  provider: 'gemini-transform',
  params: {
    prompt: 'Make the sky more vibrant with dramatic sunset colors',
  }
});

await floimg.save(edited, './edited-image.png');

Pre-Prompt for Dynamic Workflows

When receiving prompts from upstream nodes (e.g., text generation), use prePrompt to add context:

const edited = await floimg.transform({
  blob: inputImage,
  op: 'edit',
  provider: 'gemini-transform',
  params: {
    prompt: dynamicPromptFromTextNode, // e.g., "a magical forest"
    prePrompt: 'Edit this image by incorporating the following concept while preserving the original composition:',
  }
});

Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Edit instructions
prePrompt	string	No	Instructions prepended to prompt (for dynamic workflows)
model	string	No	Model (gemini-2.5-flash-image or gemini-3-pro-image-preview)
aspectRatio	string	No	Output aspect ratio (see below)
imageSize	string	No	Output resolution: 1K, 2K, or 4K (default: 1K)
groundingWithSearch	boolean	No	Enable Google Search grounding for real-time data
apiKey	string	No	Per-request API key override

Aspect Ratios

Supported values: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9

Image Sizes

Size	Resolution	Use Case
1K	~1024px	Default, fast generation
2K	~2048px	Higher quality output
4K	~4096px	Print/archival quality

Google Search Grounding

When enabled, the model can generate images based on real-time data:

const edited = await floimg.transform({
  blob: inputImage,
  op: 'edit',
  provider: 'gemini-transform',
  params: {
    prompt: 'Add the current weather forecast for Tokyo as an overlay',
    groundingWithSearch: true,
    aspectRatio: '16:9',
    imageSize: '2K',
  }
});

Models

Model	Description
gemini-2.5-flash-image	Fast, high-volume (default)
gemini-3-pro-image-preview	Higher quality, better text rendering

---

Gemini Generate

Generate images from text prompts using Gemini’s native image generation.

Registration

import createClient from '@teamflojo/floimg';
import { geminiGenerate } from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerGenerator(geminiGenerate());

Usage

const image = await floimg.generate({
  generator: 'gemini-generate',
  params: {
    prompt: 'A cozy coffee shop interior with warm lighting and plants',
    aspectRatio: '16:9',
    imageSize: '2K',
  }
});

await floimg.save(image, './coffee-shop.png');

Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Image description
model	string	No	Model (default: gemini-2.5-flash-image)
aspectRatio	string	No	Output aspect ratio (default: 1:1)
imageSize	string	No	Output resolution: 1K, 2K, or 4K (default: 1K)
groundingWithSearch	boolean	No	Enable Google Search grounding
apiKey	string	No	Per-request API key override

Aspect ratios, image sizes, and Google Search grounding work the same as in Gemini Edit.

---

Google Imagen

Generate images using Google’s Imagen model (separate from Gemini).

Registration

import createClient from '@teamflojo/floimg';
import googleImagen from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerGenerator(googleImagen());

Usage

const image = await floimg.generate({
  generator: 'google-imagen',
  params: {
    prompt: 'A photorealistic mountain lake at sunrise',
    aspectRatio: '16:9',
  }
});

Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Image description (max 480 tokens)
model	string	No	Imagen model variant
aspectRatio	string	No	1:1, 16:9, 9:16, 4:3, 3:4 (default: 1:1)
numberOfImages	number	No	Images to generate (1-4)

Imagen Models

Model	Description
imagen-4.0-generate-001	Latest Imagen 4.0 (default)
imagen-4.0-fast-generate-001	Faster variant
imagen-4.0-ultra-generate-001	Highest quality
imagen-3.0-generate-002	Imagen 3.0

---

Complete Pipeline Example

Combine text, vision, and editing in a workflow:

import createClient from '@teamflojo/floimg';
import { geminiText, geminiVision, geminiTransform } from '@teamflojo/floimg-google';

const floimg = createClient();
floimg.registerTextProvider(geminiText());
floimg.registerVisionProvider(geminiVision());
floimg.registerTransformProvider(geminiTransform());

// 1. Analyze the input image
const analysis = await floimg.analyzeImage({
  provider: 'gemini-vision',
  blob: inputImage,
  params: {
    prompt: 'Describe the mood and style of this image',
  }
});

// 2. Generate enhancement suggestions
const suggestions = await floimg.generateText({
  provider: 'gemini-text',
  params: {
    prompt: 'Based on this image analysis, suggest an edit to make it more dramatic',
    context: analysis.content,
    outputFormat: 'json',
  }
});

// 3. Apply the AI-suggested edit
const edited = await floimg.transform({
  blob: inputImage,
  op: 'edit',
  provider: 'gemini-transform',
  params: {
    prompt: suggestions.parsed.suggestion,
    prePrompt: 'Apply this enhancement while preserving the original composition:',
  }
});

await floimg.save(edited, './enhanced-image.png');

See Also

• xAI (Grok) - Alternative AI text and vision
• OpenAI - DALL-E and GPT-4 Vision
• CLI AI Commands - Command-line AI tools
• Studio AI Nodes - Visual workflow AI nodes