Cloudinary Asset Orchestrator Server

This Model Context Protocol (MCP) server furnishes Cloudinary Upload and Admin API capabilities as callable instruments for Artificial Intelligence entities. This integration allows sophisticated AI environments to initiate and interface with your specified Cloudinary media repository.

Operational Mechanics

The MCP service layer:

• Executes authorized requests against the Cloudinary backend infrastructure on your behalf.
• Grants the capability to transfer media files into Cloudinary storage.
• Facilitates the governance and modification of extant assets residing in your Cloudinary tenancy.

It fundamentally leverages the documented Cloudinary API endpoints to execute these operations. Note that comprehensive coverage of every method and parameter is ongoing; expansion is planned incrementally.

Should a specific API utility be required, please submit a feature request via an issue.

Key Advantages

• Transform standard Cloudinary operational procedures into actionable tools consumable by intelligent assistants.
• Render your existing Cloudinary media collection accessible as contextual data for AI processing.

Deployment with Claude Desktop Environment

Prerequisites

• A functional NodeJS runtime environment.
• An active MCP Client application (e.g., Claude Desktop).
• Procurement and secure storage of your Cloudinary credentials from: API KEYS

Configuration Guide

To integrate this orchestration service with the Claude Desktop application, incorporate the subsequent configuration object within the "mcpServers" block of your claude_desktop_config.json file:

{ "mcpServers": { "cloudinary-mcp-server": { "command": "npx", "args": ["-y", "cloudinary-mcp-server"], "env": { "CLOUDINARY_CLOUD_NAME": "", "CLOUDINARY_API_KEY": "", "CLOUDINARY_API_SECRET": "" } } } }

• CLOUDINARY_CLOUD_NAME: Your designated Cloudinary account identifier.
• CLOUDINARY_API_KEY: The authentication key assigned to your media environment.
• CLOUDINARY_API_SECRET: The confidential security credential for your cloud instance.

Available Operations (Tools)

The following specific functionalities are exposed:

1. upload

   • Functionality: Ingests a file (asset) into the Cloudinary platform.
   • Inputs:
     • source: Data payload (URL, local file path, Base64 representation, or raw binary content) intended for upload.
     • folder: Optional path specifying the destination folder within Cloudinary.
     • publicId: Optional user-defined identifier for the uploaded resource.
     • resourceType: Categorization of the item being uploaded (e.g., image, video, raw format, automatic detection).
     • tags: A delimited string listing tags to be associated with the asset.

2. delete-asset

   • Functionality: Eradicates a designated file (asset) from Cloudinary storage.
   • Inputs:
     • publicId: The unique public identifier of the asset designated for removal.
     • assetId: The internal asset identifier for the resource targeted for deletion.

3. get-asset

   • Functionality: Retrieves the descriptive metadata for an individual file (asset).
   • Inputs:
     • assetId: The Cloudinary system-assigned identifier for the asset.
     • publicId: The public identifier associated with the asset.
     • resourceType: The classification of the asset (image, raw, video).
     • type: The asset delivery context (e.g., upload, private, authenticated).
     • tags: Boolean flag to include the associated tag list in the response.
     • context: Boolean flag to include contextual metadata in the response.
     • metadata: Boolean flag to include structured metadata fields.

4. find-assets

   • Functionality: Executes a search query against existing Cloudinary assets based on specified criteria.
   • Inputs:
     • expression: A search predicate string (e.g., 'tags=landscape' or 'public_id:archive/*').
     • resourceType: The type of resource to search within (image, video, raw).
     • maxResults: The upper limit on the quantity of records returned (constrained between 1 and 500).
     • nextCursor: A pointer to resume retrieval from a previous paginated result set.
     • tags: Flag to include asset tags in the returned collection.
     • context: Flag to include contextual information in the returned collection.

5. get-usage

   • Functionality: Generates a comprehensive summary report detailing the consumption metrics of your product environment, encompassing storage utilization, credit balance, bandwidth throughput, API request counts, resource inventory, and add-on service utilization.
   • Inputs:
     • date: An optional parameter specifying the report date in 'yyyy-mm-dd' format. Must reference a date within the preceding three calendar months. Defaults to the present date if omitted.