cloudinary

The Cloudinary MCP Server exposes Cloudinary Upload and Admin API capabilities as tools that AI assistants can call. This allows an AI to upload assets, fetch asset details, search for assets, delete assets, and retrieve usage information from your Cloudinary account, all through structured tool calls. It turns Cloudinary cloud operations into programmable actions that can be invoked by your AI workflows or chat-based assistants.

To use it, run the MCP server (e.g., via npx) with your Cloudinary API credentials in the environment. The server then exposes a set of tools such as upload, delete-asset, get-asset, find-assets, and get-usage. Each tool accepts parameters as defined in the README, enabling automation scenarios like programmatically uploading assets, organizing assets into folders, querying metadata, or auditing usage and quotas. When integrating with an AI assistant, you provide the tool name and arguments, and the MCP server forwards the request to Cloudinary and returns results to the AI.

Prerequisites:

• Node.js and npm installed on your machine
• Access to Cloudinary account API keys (cloud name, API key, API secret)

Installation steps:

1. Install Node.js from the official website if you don’t already have it.
2. Install the MCP server package locally or run via npx as shown in the usage example.
3. Prepare Cloudinary API credentials in your environment or in your client configuration.

Example setup using npx (no local install required):

# Ensure you have Node.js installed
node -v
npm -v

# Run the MCP server via npx (as described in the README example)
npx -y cloudinary-mcp-server

Environment variables (required by the server):

export CLOUDINARY_CLOUD_NAME="<cloud name>"
export CLOUDINARY_API_KEY="<api-key>"
export CLOUDINARY_API_SECRET="<api-secret>"

If you prefer a local install, you can install the package and run the server script directly according to the package's documentation, then configure the same environment variables.

Tips and considerations:

• Ensure your Cloudinary API credentials are kept secure and not committed to code repositories.
• The available tools (upload, delete-asset, get-asset, find-assets, get-usage) support various Cloudinary resource types (image, video, raw, etc.). Review parameter options to tailor requests to your needs.
• When using find-assets, construct expressive queries and handle pagination via maxResults and nextCursor.
• If you encounter authentication errors, double-check CLOUDINARY_CLOUD_NAME, CLOUDINARY_API_KEY, and CLOUDINARY_API_SECRET values in your environment.
• The MCP server acts as a bridge to Cloudinary; latency will depend on network connectivity to Cloudinary endpoints.