🖼️ Unsplash Smart MCP Server

What is 🖼️ Unsplash Smart MCP Server?

A FastMCP server that enables AI agents to search, recommend, and deliver professional stock photos from Unsplash with context-aware image selection and automatic attribution management. It is built for AI agents such as Claude in Cursor and other MCP-compatible systems.

How to use 🖼️ Unsplash Smart MCP Server?

Install via local clone, Docker, or Smithery, then configure your Cursor MCP settings (mcp.json) with the npx tsx command, the server’s absolute path, and your UNSPLASH_ACCESS_KEY environment variable. Alternatively, use the Cursor CLI or a docker exec command. Agents invoke the single stock_photo tool with natural‑language prompts to search, receive URLs, and download images with automatic attribution.

Key features of 🖼️ Unsplash Smart MCP Server

• Single stock_photo tool unifies the entire image workflow.
• Context‑aware image selection from abstract or vague prompts.
• URL‑first approach conserves bandwidth and storage.
• Automatic attribution and metadata embedding for Unsplash compliance.
• Project‑aware file organization (Next.js, React, Vue, etc.).
• Attribution management system with JSON, HTML, and React generators.

Use cases of 🖼️ Unsplash Smart MCP Server

• AI agents in Cursor searching for hero images using natural language.
• Rapid prototyping where images are fetched and placed into the correct project folder.
• Enterprise applications needing unattributed image usage tracking via a local database.
• Teams that want to minimise tool spam and API calls when fetching stock photos.

FAQ from 🖼️ Unsplash Smart MCP Server

What makes this different from other Unsplash integrations?

It is purpose‑built for AI agents with a unified stock_photo tool, context‑aware selection, URL‑first delivery, and built‑in attribution management — reducing tool spam and developer effort.

What are the runtime requirements?

Node.js 18.x or higher and an Unsplash API access key (obtainable at unsplash.com/developers).

How does automatic attribution work?

Every downloaded image stores photographer information in a local database. Metadata is embedded into image files (EXIF, IPTC, XMP), and built‑in tools generate HTML or React attribution components on demand.

Where does image data live?

The server returns URLs rather than binary data. Images are downloaded by the developer to their project; attribution data is stored locally in a file‑based database.

Which transports and authentication are supported?

The server uses standard MCP transport via command-line I/O (stdin/stdout). Authentication is handled via the UNSPLASH_ACCESS_KEY environment variable. No TCP port or HTTP transport is used.