### Local Development Setup

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Instructions for cloning the repository, installing dependencies, building the project, and running the server locally.

```APIDOC
## Local Development

```bash
git clone https://github.com/TruCopilot/image-generator-vibe-coding.git
cd image-generator-vibe-coding
npm install
npm run build          # Compiles TypeScript → dist/

# Run the server manually for testing
OPENROUTER_API_KEY=sk-or-v1-your-key node dist/index.js
# → Image Generator MCP server running on stdio (default provider: openrouter)
```
```

--------------------------------

### Local Development Setup Commands

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Provides commands for cloning the repository, installing dependencies, building the project, and running the server locally with an API key.

```bash
git clone https://github.com/TruCopilot/image-generator-vibe-coding.git
cd image-generator-vibe-coding
npm install
npm run build          # Compiles TypeScript → dist/

# Run the server manually for testing
OPENROUTER_API_KEY=sk-or-v1-your-key node dist/index.js
# → Image Generator MCP server running on stdio (default provider: openrouter)
```

--------------------------------

### Clone and Install Project

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Commands to clone the repository, navigate into the project directory, install dependencies, and build the project.

```bash
git clone https://github.com/TruCopilot/image-generator-vibe-coding.git
cd image-generator-vibe-coding
npm install
npm run build
```

--------------------------------

### Quick Install for Project Scope

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Installs the image generator specifically for the current project scope. Replace YOUR_OPENROUTER_API_KEY with your key.

```bash
claude mcp add image-generator \
  -e OPENROUTER_API_KEY=YOUR_OPENROUTER_API_KEY \
  -- npx -y @trucopilot/image-generator-vibe-coding
```

--------------------------------

### Generate Image Examples

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Provides examples for generating images with different parameters, including model shortcuts, aspect ratios, and output directories. It also shows how to force the Gemini provider.

```text
// Generate a square avatar with the "flash" shortcut:
prompt:      "Friendly robot mascot, 3D rendered, studio lighting, white background, clean edges"
model:       "flash"
aspectRatio: "1:1"
imageSize:   "2K"
outputDir:   "./public/images/avatars"
```

```text
// Use quality model for a cinematic ultra-wide header:
prompt:      "Futuristic city skyline at night, neon reflections on wet pavement, 21:9 panorama"
model:       "google/gemini-3-pro-image-preview"
aspectRatio: "21:9"
imageSize:   "4K"
outputDir:   "./public/images/headers"
```

```text
// Force Gemini direct provider:
prompt:      "Korean bibimbap bowl, overhead shot, restaurant-grade food photography"
model:       "gemini-2.5-flash-image"
aspectRatio: "4:3"
imageSize:   "2K"
outputDir:   "./public/images/menu"
provider:    "gemini"
```

--------------------------------

### Quick Install for User Scope

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Use this command to install the image generator for your user scope. Ensure you replace YOUR_OPENROUTER_API_KEY with your actual key from openrouter.ai.

```bash
claude mcp add --scope user image-generator \
  -e OPENROUTER_API_KEY=YOUR_OPENROUTER_API_KEY \
  -- npx -y @trucopilot/image-generator-vibe-coding
```

--------------------------------

### Generate Image Prompt Example

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Example of a natural language prompt for generating an image. This can be used directly with the AI assistant.

```text
"Generate a hero image of a futuristic cityscape at night, neon lights, rain reflections"
```

--------------------------------

### Generate Banner Prompt Example

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Example of a prompt for generating a banner image with a specified aspect ratio. Useful for blog posts or headers.

```text
"Create a 16:9 banner for my blog post about machine learning"
```

--------------------------------

### Generate Avatar Prompt Example

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Example of a prompt for generating a square avatar image with specific style and lighting. Suitable for profile pictures.

```text
"Generate a 1:1 avatar of a friendly robot mascot, 3D rendered, studio lighting"
```

--------------------------------

### Edit Image Prompt Example

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Example of a natural language prompt for editing an existing image. Specify desired modifications.

```text
"Edit this image to add dramatic clouds and golden hour lighting"
```

--------------------------------

### Run Project Locally

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Command to run the project locally after installation and build. Requires setting the OPENROUTER_API_KEY environment variable.

```bash
OPENROUTER_API_KEY=your-key node dist/index.js
```

--------------------------------

### Image Editing Examples with Parameters

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Demonstrates various use cases for the `edit_image` tool, including re-cropping, restyling product photos, and adding seasonal flair.

```text
// Re-crop a portrait image to landscape for a hero banner:
prompt:      "Keep the subject, expand background to fill 16:9 landscape composition"
imagePath:   "./assets/team/ceo-portrait.jpg"
aspectRatio: "16:9"
outputDir:   "./public/images/team"

// Restyle a product photo:
prompt:      "Change the background to a clean white studio backdrop, enhance product sharpness"
imagePath:   "./uploads/product-raw.png"
model:       "pro"
outputDir:   "./public/images/products"

// Add seasonal flair to an existing asset:
prompt:      "Overlay subtle falling snowflakes and a soft winter color grade"
imagePath:   "./public/images/hero.png"
model:       "flash"
outputDir:   "./public/images/seasonal"
```

--------------------------------

### Run Development Server with Watch Mode

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Starts the development server and watches for file changes to recompile TypeScript.

```bash
npm run dev            # tsc --watch
```

--------------------------------

### Vibe Coding Prompt for Image Generation Setup

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

This prompt should be dropped into AI coding tools to automate the setup of the image-generator MCP server and CLAUDE.md rules. Remember to replace the placeholders for your OpenRouter API key and desired model name.

```plaintext
Fetch this doc: https://github.com/TruCopilot/image-generator-vibe-coding and install the
image-generator MCP server for me. Use user scope if possible, otherwise project scope.

Install command:
claude mcp add --scope user image-generator \
  -e OPENROUTER_API_KEY=<b>[YourOpenRouterAPIKey]</b> \
  -- npx -y @trucopilot/image-generator-vibe-coding

Then update my CLAUDE.md (or AGENTS.md) and your memory with these image generation rules:

# Image Generation (MANDATORY for all visual content)

- MCP Server: `image-generator` (via @trucopilot/image-generator-vibe-coding)
- Model: `<b>[YourModelName]</b>`
  - OpenRouter: `google/gemini-2.5-flash-image` (fast) or `google/gemini-3-pro-image-preview` (quality)
  - Gemini direct: `gemini-2.5-flash-image` (fast) or `gemini-3-pro-image-preview` (quality)
- Provider: OpenRouter (default)

## Rules
- ALWAYS use the `image-generator` MCP to generate images when working on new designs,
  building UI, or making the site more beautiful — hero sections, banners, cards,
  backgrounds, avatars, and any visual content
- CRITICAL: Always call via sub-agent (Agent tool) — base64 image data will crash
  the main context window if returned directly
- Resolution: Always `2K` — never go below unless I explicitly ask
- Style: Ultra-realistic, high detail, professional photography quality — include
  lighting, composition, and mood descriptors in every prompt
- Aspect ratios — choose based on design context:
  - `1:1`  → Avatars, profile pics, square cards, thumbnails
  - `16:9` → Hero banners, page headers, blog covers, landscape backgrounds
  - `9:16` → Mobile splash screens, story formats, vertical banners
  - `3:4` / `4:3` → Product cards, feature sections
  - `2:3` / `3:2` → Portrait/landscape editorial layouts
- Output directory: `./public/images/generated/` (or project-appropriate path)
- After generating, use the saved file path in &lt;img&gt; or CSS background-image
  — never embed base64 in markup

## Sub-Agent Pattern (Required)
Always generate images through a sub-agent like this:
  Agent tool → "Use the image-generator MCP generate_image tool with:
    prompt: '&lt;detailed visual description&gt;',
    model: '[YourModelName]',
    aspectRatio: '&lt;pick based on context&gt;',
    imageSize: '2K',
    outputDir: './public/images/generated/'
  Report back ONLY the saved file path, do NOT return image data."

Save this to your persistent memory so every future session uses these rules automatically.
```

--------------------------------

### Aspect Ratio & Resolution Reference

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

A guide to choosing the right aspect ratio for images to fit UI layouts, along with a reference for base resolutions and multipliers.

```APIDOC
## Aspect Ratio & Resolution Reference

Choosing the right aspect ratio is critical for images to fit naturally into UI layouts without CSS distortion.

### Ratio & Base Resolution Table
```
Ratio   Base Resolution    Best For
──────────────────────────────────────────────────────────────
1:1     1024×1024          Avatars, profile pics, square cards
16:9    1344×768           Hero banners, page headers, blog covers
9:16    768×1344           Mobile splash, story formats, vertical banners
3:4     864×1184           Product cards, feature sections (portrait)
4:3     1184×864           Feature sections, content blocks (landscape)
2:3     832×1248           Editorial portraits, magazine layouts
3:2     1248×832           Editorial landscapes
4:5     896×1152           Instagram posts, social media
5:4     1152×896           Social media landscape
21:9    1536×672           Ultra-wide cinematic headers
```

### Image Size Multipliers
```
imageSize multipliers:
  0.5K → ~512px base   1K → ~1024px base   2K → ~2048px base   4K → ~4096px base
```
```

--------------------------------

### Sub-agent Pattern for Image Editing

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Example of how a sub-agent can use the `edit_image` tool with specific parameters. The expected output format is also shown.

```text
// Sub-agent pattern for editing:
Agent tool → "Use the image-generator MCP edit_image tool with:
  prompt: 'Add dramatic storm clouds and change the lighting to a moody blue dusk atmosphere',
  imagePath: './public/images/generated/image-1719000000000-0.png',
  model: 'google/gemini-2.5-flash-image',
  aspectRatio: '16:9',
  outputDir: './public/images/generated/'
  Report back ONLY the saved file path."

// Expected output:
// Edited image using openrouter → google/gemini-2.5-flash-image
// Source: /your/project/public/images/generated/image-1719000000000-0.png
// Saved: /your/project/public/images/generated/image-1719000000001-0.png
```

--------------------------------

### Generate Image Tool Usage (Sub-agent Pattern)

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Demonstrates the recommended sub-agent pattern for invoking the `generate_image` tool within AI coding sessions. This prevents context window overflow by reporting only the file path.

```text
Tool: generate_image

Parameters:
  prompt      (string, required)  – Text description of the image
  model       (string, default: "google/gemini-2.5-flash-image")
              Accepts full OpenRouter IDs, full Gemini IDs, or shortcuts: "flash", "pro"
  aspectRatio (enum,   default: "1:1")
              Options: "1:1" | "16:9" | "9:16" | "3:4" | "4:3" | "2:3" | "3:2" | "4:5" | "5:4" | "21:9"
  imageSize   (enum,   default: "2K")
              Options: "0.5K" | "1K" | "2K" | "4K"
  outputDir   (string, default: "./generated-images")
  provider    (enum,   optional)  – "openrouter" | "gemini"  (auto-detected if omitted)
```

```text
// Sub-agent pattern (recommended inside AI coding sessions):
Agent tool → "Use the image-generator MCP generate_image tool with:
  prompt: 'Stunning 16:9 hero banner — cherry blossom avenue at golden hour,
           cinematic depth of field, soft bokeh, warm amber tones, ultra-realistic,
           professional photography',
  model: 'google/gemini-2.5-flash-image',
  aspectRatio: '16:9',
  imageSize: '2K',
  outputDir: './public/images/generated/'
  Report back ONLY the saved file path, do NOT return image data."

// Expected output (returned by the tool):
// Generated 1 image(s) using openrouter → google/gemini-2.5-flash-image
// Aspect ratio: 16:9 | Size: 2K (2048x2048)
// Saved: /your/project/public/images/generated/image-1719000000000-0.png
```

--------------------------------

### Manual JSON Configuration (Google Gemini)

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Manually configure the MCP server using JSON, specifying the Google Gemini provider by using `GEMINI_API_KEY`.

```json
{
  "mcpServers": {
    "image-generator": {
      "command": "npx",
      "args": ["-y", "@trucopilot/image-generator-vibe-coding"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-key-here"
      }
    }
  }
}
```

--------------------------------

### Environment Variables

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Lists the required environment variables for configuring API keys for OpenRouter and Google Gemini.

```APIDOC
## Environment Variables

```bash
# .env.example

# OpenRouter (preferred) — access 300+ models via one API key
OPENROUTER_API_KEY=sk-or-v1-your-key-here

# Google Gemini direct — used as fallback if OPENROUTER_API_KEY is not set
GEMINI_API_KEY=AIzaSy-your-key-here

# At least one key is required — the server exits with an error if both are missing.
```
```

--------------------------------

### Environment Variables for API Keys

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Specifies the required environment variables for authentication: `OPENROUTER_API_KEY` (preferred) and `GEMINI_API_KEY` (fallback).

```bash
# .env.example

# OpenRouter (preferred) — access 300+ models via one API key
OPENROUTER_API_KEY=sk-or-v1-your-key-here

# Google Gemini direct — used as fallback if OPENROUTER_API_KEY is not set
GEMINI_API_KEY=AIzaSy-your-key-here

# At least one key is required — the server exits with an error if both are missing.
```

--------------------------------

### Internal Model Map Configuration

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Illustrates the internal mapping of model shortcuts to their full provider-specific IDs for both OpenRouter and Gemini providers.

```typescript
// Internal MODEL_MAP (src/index.ts)
const MODEL_MAP = {
  flash: {
    openrouter: "google/gemini-2.5-flash-image",
    gemini:     "gemini-2.5-flash-image",
  },
  pro: {
    openrouter: "google/gemini-3-pro-image-preview",
    gemini:     "gemini-3-pro-image-preview",
  },
  // Full IDs also accepted as keys — resolved to the correct provider variant
};

// Any OpenRouter model ID not in the map is passed through unchanged,
enabling access to 300+ models listed at https://openrouter.ai/models
// Example — use any third-party image model:
prompt:   "Impressionist oil painting of a Parisian café"
model:    "stability-ai/stable-diffusion-3-5-large"   // any OpenRouter model ID
provider: "openrouter"
```

--------------------------------

### Edit Image Tool Parameters

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Defines the parameters for the `edit_image` tool, including prompt, image path, model, aspect ratio, output directory, and provider.

```text
Tool: edit_image

Parameters:
  prompt      (string, required)  – Edit instructions
  imagePath   (string, required)  – Absolute or relative path to the source image
  model       (string, default: "google/gemini-2.5-flash-image")
  aspectRatio (enum,   optional)  – Output aspect ratio (changes crop/composition)
  outputDir   (string, default: "./generated-images")
  provider    (enum,   optional)  – "openrouter" | "gemini"
```

--------------------------------

### generate_image MCP Tool

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Generates a new image from a text prompt using the specified model and provider. The result is auto-saved as a timestamped PNG file in the output directory, and the saved file path is returned alongside a preview of the image data.

```APIDOC
## MCP Tool: `generate_image`

### Description
Generates a new image from a text prompt using the specified model and provider. The result is auto-saved as a timestamped PNG file in the output directory, and the saved file path is returned alongside a preview of the image data. When called from an AI assistant, this tool should always be invoked via a sub-agent to prevent base64 image data from overflowing the main context window.

### Parameters
#### Required Parameters
- **prompt** (string) - Text description of the image

#### Optional Parameters
- **model** (string, default: "google/gemini-2.5-flash-image") - Accepts full OpenRouter IDs, full Gemini IDs, or shortcuts: "flash", "pro"
- **aspectRatio** (enum, default: "1:1") - Options: "1:1" | "16:9" | "9:16" | "3:4" | "4:3" | "2:3" | "3:2" | "4:5" | "5:4" | "21:9"
- **imageSize** (enum, default: "2K") - Options: "0.5K" | "1K" | "2K" | "4K"
- **outputDir** (string, default: "./generated-images") - The directory to save generated images.
- **provider** (enum, optional) - "openrouter" | "gemini" (auto-detected if omitted)

### Request Example (Sub-agent pattern)
```
Agent tool → "Use the image-generator MCP generate_image tool with:
  prompt: 'Stunning 16:9 hero banner — cherry blossom avenue at golden hour,
           cinematic depth of field, soft bokeh, warm amber tones, ultra-realistic,
           professional photography',
  model: 'google/gemini-2.5-flash-image',
  aspectRatio: '16:9',
  imageSize: '2K',
  outputDir: './public/images/generated/'
  Report back ONLY the saved file path, do NOT return image data."
```

### Response Example
```
// Generated 1 image(s) using openrouter → google/gemini-2.5-flash-image
// Aspect ratio: 16:9 | Size: 2K (2048x2048)
// Saved: /your/project/public/images/generated/image-1719000000000-0.png
```

### Examples
```
// Generate a square avatar with the "flash" shortcut:
prompt:      "Friendly robot mascot, 3D rendered, studio lighting, white background, clean edges"
model:       "flash"
aspectRatio: "1:1"
imageSize:   "2K"
outputDir:   "./public/images/avatars"

// Use quality model for a cinematic ultra-wide header:
prompt:      "Futuristic city skyline at night, neon reflections on wet pavement, 21:9 panorama"
model:       "google/gemini-3-pro-image-preview"
aspectRatio: "21:9"
imageSize:   "4K"
outputDir:   "./public/images/headers"

// Force Gemini direct provider:
prompt:      "Korean bibimbap bowl, overhead shot, restaurant-grade food photography"
model:       "gemini-2.5-flash-image"
aspectRatio: "4:3"
imageSize:   "2K"
outputDir:   "./public/images/menu"
provider:    "gemini"
```
```

--------------------------------

### Manual JSON Configuration for MCP Servers

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Configure the image-generator MCP server manually using a JSON file. This is compatible with tools like Claude Code, Cursor, and Windsurf. Replace 'your-openrouter-key-here' with your API key.

```json
{
  "mcpServers": {
    "image-generator": {
      "command": "npx",
      "args": ["-y", "@trucopilot/image-generator-vibe-coding"],
      "env": {
        "OPENROUTER_API_KEY": "your-openrouter-key-here"
      }
    }
  }
}
```

--------------------------------

### generate_image

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Generates an image from a text prompt using various models and configurations.

```APIDOC
## `generate_image`

### Description
Generate an image from a text prompt.

### Parameters
#### Path Parameters
- `prompt` (string) - Required - Text description of the image
- `model` (string) - Optional - `google/gemini-2.5-flash-image` - Full model ID — see [Models](#-models). Shortcuts: `flash`, `pro`
- `aspectRatio` (enum) - Optional - `1:1` - `1:1` · `16:9` · `9:16` · `3:4` · `4:3` · `2:3` · `3:2` · `4:5` · `5:4` · `21:9`
- `imageSize` (enum) - Optional - `2K` - `0.5K` · `1K` · `2K` · `4K`
- `outputDir` (string) - Optional - `./generated-images` - Directory to save generated images
- `provider` (enum) - Optional - auto-detect - `openrouter` · `gemini`
```

--------------------------------

### edit_image Tool

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Edits an existing image on disk by applying text-based instructions. Supports various image formats and allows customization of model, aspect ratio, output directory, and provider.

```APIDOC
## Tool: edit_image

### Description
Edits an existing image on disk by applying text-based instructions using the same Gemini-backed providers. The source image is read, base64-encoded, and sent alongside the edit prompt. Supports `.png`, `.jpg`, `.jpeg`, `.webp`, and `.gif` inputs. The edited result is saved as a new PNG in the output directory; the original file is never overwritten.

### Parameters
- **prompt** (string, required) – Edit instructions
- **imagePath** (string, required) – Absolute or relative path to the source image
- **model** (string, default: "google/gemini-2.5-flash-image")
- **aspectRatio** (enum, optional) – Output aspect ratio (changes crop/composition)
- **outputDir** (string, default: "./generated-images")
- **provider** (enum, optional) – "openrouter" | "gemini"

### Usage Examples

#### Sub-agent pattern for editing:
```
Agent tool → "Use the image-generator MCP edit_image tool with:
  prompt: 'Add dramatic storm clouds and change the lighting to a moody blue dusk atmosphere',
  imagePath: './public/images/generated/image-1719000000000-0.png',
  model: 'google/gemini-2.5-flash-image',
  aspectRatio: '16:9',
  outputDir: './public/images/generated/'
  Report back ONLY the saved file path."
```

#### Expected output:
```
Edited image using openrouter → google/gemini-2.5-flash-image
Source: /your/project/public/images/generated/image-1719000000000-0.png
Saved: /your/project/public/images/generated/image-1719000000001-0.png
```

#### Re-crop a portrait image to landscape for a hero banner:
```
prompt:      "Keep the subject, expand background to fill 16:9 landscape composition"
imagePath:   "./assets/team/ceo-portrait.jpg"
aspectRatio: "16:9"
outputDir:   "./public/images/team"
```

#### Restyle a product photo:
```
prompt:      "Change the background to a clean white studio backdrop, enhance product sharpness"
imagePath:   "./uploads/product-raw.png"
model:       "pro"
outputDir:   "./public/images/products"
```

#### Add seasonal flair to an existing asset:
```
prompt:      "Overlay subtle falling snowflakes and a soft winter color grade"
imagePath:   "./public/images/hero.png"
model:       "flash"
outputDir:   "./public/images/seasonal"
```
```

--------------------------------

### Provider & Model Resolution

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Details on how the server resolves the active provider and model ID, including environment variable precedence and model mapping.

```APIDOC
## Provider & Model Resolution

The server resolves the active provider and model ID at request time. `OPENROUTER_API_KEY` takes precedence; if only `GEMINI_API_KEY` is set, the Gemini provider is used. Model shortcuts are mapped to their full provider-specific IDs internally.

### Internal MODEL_MAP
```typescript
const MODEL_MAP = {
  flash: {
    openrouter: "google/gemini-2.5-flash-image",
    gemini:     "gemini-2.5-flash-image",
  },
  pro: {
    openrouter: "google/gemini-3-pro-image-preview",
    gemini:     "gemini-3-pro-image-preview",
  },
  // Full IDs also accepted as keys — resolved to the correct provider variant
};
```

Any OpenRouter model ID not in the map is passed through unchanged, enabling access to 300+ models listed at https://openrouter.ai/models.

### Example using a third-party model:
```
prompt:   "Impressionist oil painting of a Parisian café"
model:    "stability-ai/stable-diffusion-3-5-large"   // any OpenRouter model ID
provider: "openrouter"
```
```

--------------------------------

### edit_image

Source: https://github.com/trucopilot/image-generator-vibe-coding/blob/main/README.md

Edits an existing image based on text instructions.

```APIDOC
## `edit_image`

### Description
Edit an existing image with text instructions.

### Parameters
#### Path Parameters
- `prompt` (string) - Required - Edit instructions
- `imagePath` (string) - Required - Path to source image
- `model` (string) - Optional - `google/gemini-2.5-flash-image` - Full model ID — same options as above
- `aspectRatio` (enum) - Optional - Output aspect ratio
- `outputDir` (string) - Optional - `./generated-images` - Directory to save edited images
- `provider` (enum) - Optional - auto-detect - `openrouter` · `gemini`
```

--------------------------------

### CLAUDE.md Image Generation Rules

Source: https://context7.com/trucopilot/image-generator-vibe-coding/llms.txt

Defines mandatory rules for using the image generator MCP for all visual content within AI coding sessions. Ensures consistent style, resolution, and output paths.

```markdown
# Image Generation (MANDATORY for all visual content)

- MCP Server: `image-generator` (via @trucopilot/image-generator-vibe-coding)
- Model: `google/gemini-2.5-flash-image` (fast) or `google/gemini-3-pro-image-preview` (quality)
- Provider: OpenRouter (default)

## Rules
- ALWAYS use the `image-generator` MCP when building UI, hero sections, banners,
  cards, backgrounds, avatars, or any visual content
- CRITICAL: Always call via sub-agent (Agent tool) — never return base64 data directly
- Resolution: Always `2K` unless explicitly asked otherwise
- Style: Ultra-realistic, professional photography quality — include lighting,
  composition, and mood descriptors in every prompt
- Output directory: `./public/images/generated/`
- After generating, use the saved file path in <img> or CSS background-image

## Sub-Agent Pattern (Required)
Agent tool → "Use the image-generator MCP generate_image tool with:
  prompt: '<detailed visual description>',
  model: 'google/gemini-2.5-flash-image',
  aspectRatio: '<pick based on context>',
  imageSize: '2K',
  outputDir: './public/images/generated/'
  Report back ONLY the saved file path, do NOT return image data."
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.