Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://www.npmjs.com/package/@drumnation/unsplash-smart-mcp-server below:

@drumnation/unsplash-smart-mcp-server - npm

Empower your AI agents with stunning visuals, zero hassle.

A powerful FastMCP server that enables AI agents to seamlessly search, recommend, and deliver professional stock photos from Unsplash with intelligent context awareness and automated attribution management.

In the landscape of visual content integration, our Unsplash Smart MCP Server stands out as the definitive solution for AI-powered image acquisition:

• 🧠 AI-Agent Optimized: Purpose-built for AI agents like Claude in Cursor, streamlining image requests with natural language
• 🔍 Context-Aware Image Selection: Interprets vague requests intelligently, delivering relevant images even from abstract prompts
• ⚡ Single Tool Efficiency: Eliminates tool spam with a unified stock_photo tool that handles the entire image workflow
• 📊 Resource Optimization: URL-first approach conserves bandwidth and storage while maintaining flexibility
• ✅ Automatic Attribution: Built-in compliance with Unsplash's Terms of Service with zero developer effort
• 📁 Project-Aware Organization: Intelligently organizes images based on your project structure (Next.js, React, Vue, etc.)
• 🧩 Seamless Integration: Designed for minimal setup and maximum compatibility with your existing workflow

• Smart Contextual Search: Find the perfect image through natural language requests
• Automatic Subject Selection: AI determines optimal image subjects from your purpose description
• Intent-Driven Results: Get images that match not just keywords, but the underlying intent
• Seamless Agent Integration: Works out-of-the-box with Claude in Cursor and other MCP-compatible agents

• Two-Step Workflow: Get URLs for controlled downloads, avoiding permission issues and unnecessary storage
• Project-Aware File Management: Auto-organizes images based on framework conventions
• Intelligent Directory Creation: Creates appropriate folder structures based on your project type
• Progressive Enhancement: Works with any project size, from quick prototypes to enterprise applications

• Complete Attribution Management:
  • Local attribution database tracks all image usage
  • Automatic embedding of photographer metadata in images (EXIF, IPTC, XMP)
  • One-click generation of attribution pages in multiple formats
  • Comprehensive API for attribution data

• Node.js 18.x or higher
• An Unsplash API access key (get one here)

1. Clone the repository:

git clone https://github.com/drumnation/unsplash-smart-mcp-server.git
cd unsplash-smart-mcp-server

1. Install dependencies:

1. Configure your Cursor MCP settings:

   • macOS: Edit ~/.cursor/mcp.json
   • Windows: Edit %USERPROFILE%\.cursor\mcp.json
   • Linux: Edit ~/.cursor/mcp.json

2. Add the following configuration:

{
  "servers": {
    "unsplash": {
      "command": "npx",
      "args": ["tsx", "src/server.ts"],
      "cwd": "/absolute/path/to/unsplash-smart-mcp-server",
      "env": {
        "UNSPLASH_ACCESS_KEY": "your_api_key_here"
      }
    }
  }
}

1. Replace:

   • /absolute/path/to/unsplash-smart-mcp-server with the actual path where you cloned the repo
   • your_api_key_here with your Unsplash API key

2. Save the file and restart Cursor.

Important: Unlike many MCP servers, this server requires direct process piping and cannot be accessed via TCP ports or through npm directly due to how it handles FastMCP's I/O interactions. The local installation method is the most reliable approach.

If you prefer using Cursor's CLI:

claude mcp add unsplash npx tsx /path/to/unsplash-smart-mcp-server/src/server.ts --cwd /path/to/unsplash-smart-mcp-server
claude mcp config set unsplash UNSPLASH_ACCESS_KEY=your_api_key_here

Replace the paths and API key with your actual values.

1. Clone the repository:

git clone https://github.com/drumnation/unsplash-smart-mcp-server.git
cd unsplash-smart-mcp-server

1. Create a docker-compose.yml file:

services:
  unsplash-mcp:
    build: .
    image: unsplash-mcp-server
    restart: always
    stdin_open: true
    tty: true
    environment:
      - UNSPLASH_ACCESS_KEY=your_api_key_here

1. Build and start the container:

1. Configure your Cursor MCP settings:

   • macOS: Edit ~/.cursor/mcp.json
   • Windows: Edit %USERPROFILE%\.cursor\mcp.json
   • Linux: Edit ~/.cursor/mcp.json

2. Add the following configuration:

{
  "servers": {
    "unsplash": {
      "command": "docker",
      "args": ["exec", "-i", "unsplash-mcp-unsplash-mcp-1", "tsx", "src/server.ts"],
      "env": {}
    }
  }
}

1. Save the file and restart Cursor.

This setup will:

• Start the server automatically when Docker starts
• Restart the server if it crashes
• Run in the background without terminal windows
• Provide a reliable connection to Cursor

If you prefer cloud deployment, you can use Smithery:

1. Install the server in Cursor via Smithery:

npx @smithery/cli install @drumnation/unsplash-smart-mcp-server --client cursor --key your_api_key_here

1. Alternatively, you can log in to Smithery.ai and deploy it through their web interface.

Note for Windows users: Smithery deployment includes special handling for Windows compatibility.

For detailed instructions and troubleshooting, see the Smithery Deployment Guide.

Our Unsplash Smart MCP Server is designed to make image acquisition through AI agents effortless and intuitive:

1. Initiate a request: Simply ask Claude for an image in natural language
2. AI interpretation: Claude understands your needs and calls the stock_photo tool with optimized parameters
3. Smart image selection: The server interprets context and finds the most relevant images
4. Presentation of options: Claude presents you with the best matches and download commands
5. Seamless download: Execute the suggested commands to place images exactly where you need them
6. Automatic attribution: All attribution data is stored and can be accessed whenever needed

This process eliminates the traditional workflow of:

1. Searching Unsplash manually
2. Scrolling through hundreds of results
3. Downloading images to random locations
4. Moving files to the correct project folders
5. Manually tracking attribution data
6. Creating attribution pages

Ask Claude in Cursor for images using natural language prompts like these:

"Find a professional image for a tech startup landing page hero section"

If you're using Windows and experiencing the "Client closed" error when running the MCP server in Cursor, follow these special configuration steps:

Create a file named mcp.json in your .cursor directory (typically at %USERPROFILE%\.cursor\mcp.json) with one of these configurations:

{
  "mcpServers": {
    "stock_photo": {
      "command": "node",
      "args": ["./node_modules/.bin/tsx", "path/to/unsplash-mcp/src/server.ts"],
      "disabled": false,
      "env": {
        "UNSPLASH_ACCESS_KEY": "your_api_key_here"
      },
      "shell": false
    }
  }
}

{
  "mcpServers": {
    "stock_photo": {
      "command": "powershell",
      "args": ["-Command", "npx tsx path/to/unsplash-mcp/src/server.ts"],
      "disabled": false,
      "env": {
        "UNSPLASH_ACCESS_KEY": "your_api_key_here"
      }
    }
  }
}

For complete documentation on Windows compatibility, see Windows Compatibility Guide.

Our architecture uses a URL-first approach rather than direct image embedding for several critical reasons:

1. Storage Efficiency: Prevents AI agents from unnecessarily storing large binary data in their context
2. Bandwidth Conservation: Reduces data transfer between services, improving response times
3. Placement Flexibility: Allows developers to download images exactly where they're needed
4. Permission Management: Avoids filesystem permission issues in restricted environments
5. Workflow Integration: Seamlessly integrates with existing development pipelines

This strategy enables AI agents to intelligently suggest the optimal download location based on project context, without being constrained by their own environment limitations.

Unlike other solutions that require multiple tool calls for searching, filtering, downloading, and attributing images, our server:

• Unifies the entire image workflow into a single stock_photo tool
• Optimizes result retrieval by requesting more images upfront to enable better filtering
• Eliminates ping-pong interactions between the agent and services
• Reduces agent token usage by streamlining request and response formats

This design significantly reduces the number of API calls and tool invocations, leading to faster results and lower operational costs.

Using images from Unsplash requires adherence to their Terms of Service. Our server handles this automatically:

1. Attribution Data Capture: Every image download automatically stores photographer information
2. Metadata Embedding: Photographer details are embedded directly into image files
3. Attribution Database: A local database maintains a record of all image usage
4. Attribution Generators: Built-in tools create HTML and React attribution components
5. API Access: Simple endpoints to retrieve attribution data for any project

By using our Unsplash Smart MCP Server, you are automatically compliant with Unsplash's requirements without any additional effort.

The server includes a comprehensive attribution management system:

// Retrieve attribution data for your project
const attributions = await fetch('http://localhost:3000/api/unsplash', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    method: 'get_attributions',
    params: {
      format: 'json',  // Options: json, html, react
      projectPath: '/path/to/your/project'
    }
  })
}).then(res => res.json());

// attributions contains complete data about every image used

The API can generate three types of attribution files:

1. JSON: Structured data for custom implementations
2. HTML: Ready-to-use HTML page for website footer or credits section
3. React: Drop-in React component for modern web applications

Our Unsplash Smart MCP Server seamlessly integrates into your development workflow:

• Instantly populate mockups with relevant placeholder images
• Maintain consistent image dimensions across components
• Organize images logically within your project structure

• Enhance technical documentation with explanatory visuals
• Create visually appealing tutorials and guides
• Maintain proper attribution for all visual assets

• Quickly find images for blog posts and articles
• Generate visuals for social media content
• Access consistent imagery for product marketing

• Populate e-commerce sites with product imagery
• Create visually rich user experiences
• Maintain separate image collections for different sections

Images are automatically organized based on your project type:

Contributions are welcome! Please feel free to submit a Pull Request.

1. Clone the repository
2. Install dependencies with npm install
3. Create a .env file with your Unsplash API key
4. Run in development mode with npm run dev
5. Run tests with npm test

Here's what we're planning for future releases:

• Image Editing Capabilities: Basic resizing, cropping, and adjustment tools
• Advanced Search Filters: More granular control over image selection
• Batch Processing: Handle multiple image requests efficiently
• Custom Collections: Save and manage groups of images for projects
• Team Collaboration: Share attribution and image collections
• Usage Analytics: Track image usage across projects
• Additional Image Sources: Integration with other stock photo providers
• Improved Timeout Handling: Enhanced timeout configuration and recovery mechanisms

MIT License

When using images from Unsplash, you must comply with the Unsplash License:

• Attribution is not required but appreciated
• You cannot sell unaltered copies of the photos
• You cannot compile photos from Unsplash to create a competing service

Our server's attribution system makes it easy to provide proper credit to photographers.

For issues or questions, please open an issue on GitHub.

# Clone the repository
git clone https://github.com/drumnation/unsplash-smart-mcp-server.git
cd unsplash-smart-mcp-server

# Install dependencies
npm install

# Set up your environment variables
cp .env.example .env
# Edit .env to add your UNSPLASH_ACCESS_KEY

# Start the development server
npm run dev

The package includes a comprehensive test suite:

# Run core tests
npm test

# Run all tests and get a summary report
npm run test:all

The test suite includes:

• Unit and integration tests
• Manual tool testing
• Docker container tests
• Smithery.ai integration tests

For detailed information about testing, see docs/testing.md.

Empower your AI agents with the perfect images, every time.
Built with ❤️ for developers and AI enthusiasts.

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4