Cloudinary Asset Management

Cloudinary Asset Management MCP Server

Summary

Table of Contents

• Cloudinary Asset Management MCP Server
  • Installation
  • Configuration
  • Authentication
  • Available Tools
  • Custom Tools
  • Progressive Discovery
• Development
  • Building from Source
  • Contributions

Installation

{
  "command": "npx",
  "args": [
    "@cloudinary/asset-management-mcp",
    "start",
    "--api-key",
    "",
    "--api-secret",
    "",
    "--cloud-name",
    ""
  ]
}

claude mcp add CloudinaryAssetMgmt -- npx -y @cloudinary/asset-management-mcp start --api-key  --api-secret  --cloud-name 

gemini mcp add CloudinaryAssetMgmt -- npx -y @cloudinary/asset-management-mcp start --api-key  --api-secret  --cloud-name 

{
  "command": "npx",
  "args": [
    "@cloudinary/asset-management-mcp",
    "start",
    "--api-key",
    "",
    "--api-secret",
    "",
    "--cloud-name",
    ""
  ]
}

{
  "command": "npx",
  "args": [
    "@cloudinary/asset-management-mcp",
    "start",
    "--api-key",
    "",
    "--api-secret",
    "",
    "--cloud-name",
    ""
  ]
}

npx @cloudinary/asset-management-mcp start --api-key  --api-secret  --cloud-name 

npx @cloudinary/asset-management-mcp --help

Configuration

Environment Variables

The MCP server supports the following environment variables:

CLOUDINARY_URL Format

You can use a single CLOUDINARY_URL instead of individual variables:

CLOUDINARY_URL=cloudinary://API_KEY:API_SECRET@CLOUD_NAME

Response Header Collection

You can configure the server to include Cloudinary API response headers (such as x-request-id and rate limit info) in tool output. This is useful for debugging and monitoring.

Set CLOUDINARY_COLLECT_HEADERS to control which headers are collected:

# Collect all response headers
CLOUDINARY_COLLECT_HEADERS=true

# Collect specific headers by exact name (comma-separated)
CLOUDINARY_COLLECT_HEADERS=x-request-id,x-featureratelimit-limit,x-featureratelimit-remaining

# Mix exact names, prefix matching, and regex matching
CLOUDINARY_COLLECT_HEADERS=x-request-id,prefix:x-featureratelimit-

Header matching specs

Each entry in the comma-separated list is matched against response header names:

You can also set this via the CLOUDINARY_URL query parameter:

CLOUDINARY_URL=cloudinary://API_KEY:API_SECRET@CLOUD_NAME?collect_headers=true

When enabled, collected headers appear in an _headers field in the tool response. When not set, no headers are collected and responses are unchanged.

MCP Apps

The server can expose interactive MCP UI Apps (spec-aligned with io.modelcontextprotocol/ui) that hosts can render alongside tool results — for example, an asset gallery for list results, a single-asset detail view, and an upload UI.

Apps are enabled by default. Use the --mcp-apps flag (available on both start and serve) or the CLOUDINARY_MCP_APPS environment variable to override this behaviour:

Available app names: asset-gallery, asset-details, asset-upload.

# Disable all apps
npx @cloudinary/asset-management-mcp serve --mcp-apps none

# Enable only the gallery
npx @cloudinary/asset-management-mcp start --mcp-apps asset-gallery

# Disable via env var
CLOUDINARY_MCP_APPS=none npx @cloudinary/asset-management-mcp start

Precedence: CLI flag > environment variable > built-in default.

Authentication

The MCP server uses your Cloudinary API key and secret for authentication:

{
  "env": {
    "CLOUDINARY_CLOUD_NAME": "demo",
    "CLOUDINARY_API_KEY": "123456789012345",
    "CLOUDINARY_API_SECRET": "abcdefghijklmnopqrstuvwxyz12"
  }
}

Available Tools

The MCP server exposes Cloudinary's Asset Management API as tools. Use your AI application to discover and invoke the available tools for uploading, managing, searching, and transforming your media assets.

Usage Examples

Example 1: Upload and Transform an Image

1. Upload a local image: "Upload file:///Users/me/photo.jpg to Cloudinary as 'hero-image'"
2. Transform it: "Transform asset 'hero-image' with transformations 'c_fill,w_800,h_600/e_sharpen'"
3. Get details: "Show me details for asset with ID [asset-id]"

Example 2: Search and Organize Assets

1. Search for images: "Find all images with tag 'product' uploaded in the last 7 days"
2. Create folder: "Create a new folder called 'summer-2024-products'"
3. List assets: "Show me all video assets in the 'marketing' folder"

Example 3: Generate Archive

1. Get transformation docs: "Show me the Cloudinary transformation reference"
2. Apply transformations: "Transform 'banner' asset with 'c_scale,w_1200/f_auto,q_auto'"
3. Create archive: "Generate a ZIP archive of all images with tag 'export-ready'"

Example 4: Asset Management Workflow

1. Upload multiple files: "Upload all images from folder /assets/new-products/"
2. Add tags: "Update asset [asset-id] and add tags 'featured,homepage'"
3. Get usage stats: "Show my Cloudinary account usage statistics"

Custom Tools

This MCP server includes two powerful custom tools:

get-tx-reference

Retrieves the complete Cloudinary transformation reference documentation.

When to use:

• Before creating or modifying transformations
• When user asks about image/video effects, resizing, cropping, filters

Example:

Use get-tx-reference to learn about available transformations

transform-asset

Applies transformations to existing assets using Cloudinary's explicit API.

Parameters:

• publicId - The asset's public ID
• transformations - Transformation string (e.g., c_fill,w_300,h_200)
• resourceType - Type: image, video, or raw (default: image)
• invalidate - Invalidate CDN cache (default: false)

Example:

Transform asset "sample" with transformations "c_fill,w_500,h_500/e_sepia"

Progressive Discovery

MCP servers with many tools can bloat LLM context windows, leading to increased token usage and tool confusion. Dynamic mode solves this by exposing only a small set of meta-tools that let agents progressively discover and invoke tools on demand.

To enable dynamic mode, pass the --mode dynamic flag when starting your server:

{
  "mcpServers": {
    "CloudinaryAssetMgmt": {
      "command": "npx",
      "args": ["@cloudinary/asset-management-mcp", "start", "--mode", "dynamic"],
      // ... other server arguments
    }
  }
}

In dynamic mode, the server registers only the following meta-tools instead of every individual tool:

• list_tools: Lists all available tools with their names and descriptions.
• describe_tool_input: Returns the input schema for one or more tools by name.
• execute_tool: Executes a tool by name with its arguments.
• list_scopes: Lists the scopes available on the server.

This approach significantly reduces the number of tokens sent to the LLM on each request, which is especially useful for servers with a large number of tools.

You can combine dynamic mode with scope and tool filters:

{
  "mcpServers": {
    "CloudinaryAssetMgmt": {
      "command": "npx",
      "args": ["@cloudinary/asset-management-mcp", "start", "--mode", "dynamic", "--scope", "admin"],
      // ... other server arguments
    }
  }
}

Development

Building from Source

Prerequisites

• Node.js v20 or higher
• npm, pnpm, bun, or yarn

Build Steps

# Clone the repository
git clone https://github.com/cloudinary/asset-management-mcp.git
cd asset-management-mcp

# Install dependencies
npm install

# Build the project
npm run build

# Run locally
npm start

Project Structure

asset-management-mcp/
├── src/
│   ├── hooks/              # SDK hooks (manual)
│   │   ├── cloudinaryAuthHook.ts   # Auth & file:// handling
│   │   ├── customHeadersHook.ts    # Inject custom request headers
│   │   ├── responseHeadersHook.ts  # Collect response headers
│   │   ├── userAgentHook.ts        # Build User-Agent string
│   │   └── registration.ts         # Hook registration
│   ├── mcp-server/         # MCP server implementation
│   │   ├── server.ts             # Main server (auto-generated)
│   │   ├── server.extensions.ts  # Custom tools & app wiring (manual)
│   │   ├── tools/                # Generated tool wrappers
│   │   └── apps/                 # MCP UI Apps (manual)
│   │       ├── config.ts             # App registry & --mcp-apps parsing
│   │       ├── cli-flag.ts           # stricli flag definition
│   │       ├── extensions.ts         # Resource-template registration
│   │       ├── uri.ts                # App URI helpers / tool-name injection
│   │       ├── tool-hooks.ts         # Per-tool app hooks
│   │       ├── app-shared.ts         # Shared app utilities
│   │       ├── asset-gallery-app.ts  # List results gallery UI
│   │       ├── asset-details-app.ts  # Single-asset detail UI
│   │       └── asset-upload-app.ts   # Upload UI
│   ├── funcs/              # API function implementations
│   └── models/             # Type definitions
├── .github/
│   └── workflows/          # CI/CD workflows
└── .speakeasy/             # Speakeasy configuration

Contributions

While we value contributions to this MCP Server, most of the code is generated programmatically from the Cloudinary API spec. Any manual changes to generated files will be overwritten on the next generation — please direct your changes to the manual extension points below.

What you can contribute:

• Custom tools and server wiring in src/mcp-server/server.extensions.ts
• MCP UI Apps in src/mcp-server/apps/ (gallery, details, upload, and new apps)
• SDK hooks in src/hooks/ (auth, custom headers, response headers, user agent)
• Documentation improvements (this README, JSDoc on manual files)
• Bug reports and feature requests

Generated files (do not edit):

• src/mcp-server/server.ts
• src/mcp-server/tools/*.ts
• src/funcs/*.ts
• src/models/*.ts

When touching generated files is unavoidable, prefer updating the upstream spec or Speakeasy configuration in .speakeasy/ so the change survives regeneration.

We look forward to hearing your feedback. Feel free to open a PR or issue with a proof of concept and we'll do our best to include it in a future release.

MCP Server Created by Speakeasy