@teamflojo/floimg-mcp

Model Context Protocol (MCP) server for FloImg - AI image workflow automation.

This package exposes FloImg's image generation, transformation, and saving capabilities to AI agents like Claude, GPT, and others through the Model Context Protocol.

Installation

npm install -g @teamflojo/floimg-mcp
# or
pnpm add -g @teamflojo/floimg-mcp

Usage

With Claude Desktop

Add to your Claude Desktop configuration (~/.claude/claude_desktop_config.json):

{
  "mcpServers": {
    "floimg": {
      "command": "npx",
      "args": ["-y", "@teamflojo/floimg-mcp"],
      "env": {
        "OPENAI_API_KEY": "sk-..."
      }
    }
  }
}

With Claude Code

Add to your project's .mcp.json:

{
  "mcpServers": {
    "floimg": {
      "command": "npx",
      "args": ["-y", "@teamflojo/floimg-mcp"]
    }
  }
}

Direct Execution

npx @teamflojo/floimg-mcp

Available Tools

generate_image

Generate any type of image with smart intent routing:

• AI Images: DALL-E, Stability AI, Replicate
• Charts: Bar, line, pie, radar (via QuickChart)
• Diagrams: Flowcharts, sequence diagrams (via Mermaid)
• QR Codes: Any URL or text
• Screenshots: Capture any webpage
• Shapes: SVG gradients and basic shapes

Intent: "golden retriever in a sunny field" → Routes to DALL-E
Intent: "QR code for https://floimg.com" → Routes to QR generator
Intent: "bar chart of sales data" → Routes to QuickChart

transform_image

Apply deterministic, pixel-precise transforms:

• resize - Exact dimensions
• blur, sharpen - Image effects
• grayscale, modulate - Color adjustments
• addText, addCaption - Text overlays
• roundCorners - Border radius
• convert - Format conversion (PNG, JPEG, WebP, AVIF)

Unlike AI regeneration, these operations modify exactly what you specify.

save_image

Save to filesystem or cloud storage:

• ./output.png → Local filesystem
• s3://bucket/key.png → Amazon S3
• r2://bucket/key.png → Cloudflare R2
• tigris://bucket/key.png → Tigris

run_pipeline

Execute multi-step workflows atomically:

{
  "steps": [
    { "generate": { "intent": "mountain landscape at sunset" } },
    { "transform": { "operation": "resize", "params": { "width": 1200, "height": 630 } } },
    { "transform": { "operation": "addCaption", "params": { "text": "Adventure Awaits" } } },
    { "save": { "destination": "s3://my-bucket/hero.png" } }
  ]
}

analyze_image

AI vision analysis with multiple providers:

• OpenAI GPT-4V
• Anthropic Claude Vision
• Google Gemini Vision
• Ollama LLaVA (local)

generate_text

AI text generation for creating prompts, descriptions, and more.

Plugin Discovery

The MCP server automatically discovers and loads installed FloImg plugins:

• @teamflojo/floimg-quickchart - Chart generation
• @teamflojo/floimg-mermaid - Diagram generation
• @teamflojo/floimg-qr - QR code generation
• @teamflojo/floimg-d3 - Data visualization
• @teamflojo/floimg-screenshot - Web screenshots
• @teamflojo/floimg-google - Gemini AI providers

Install plugins globally for the MCP server to find them:

npm install -g @teamflojo/floimg-quickchart @teamflojo/floimg-mermaid @teamflojo/floimg-qr

Configuration

Configure via environment variables or floimg.config.ts:

# AI Providers
OPENAI_API_KEY=sk-...
GOOGLE_AI_API_KEY=...
ANTHROPIC_API_KEY=...

# Cloud Storage
AWS_ACCESS_KEY_ID=...
AWS_SECRET_ACCESS_KEY=...
AWS_REGION=us-east-1

Requirements

• Node.js 18+
• @teamflojo/floimg (peer dependency)

Related

• FloImg Documentation
• FloImg Studio - Visual workflow builder
• Model Context Protocol

License

MIT