Englisch | 简体中文
Ein einfacher MCP-Server für die nahtlose Unsplash-Bildintegration und Suchfunktionen.
Der Unsplash MCP-Server wird verwendet, um nach umfangreichen, hochwertigen Bildern zu suchen. Er ist ideal für Entwickler, die Unsplash-Funktionen in ihre eigenen Anwendungen integrieren möchten.
Related MCP server: Unsplash MCP Server
Erweiterte Bildsuche: Durchsuchen Sie die umfangreiche Fotobibliothek von Unsplash mit Filtern für:
Schlüsselwortrelevanz
Farbschemata
Ausrichtungsoptionen
Benutzerdefinierte Sortierung und Paginierung
Bevor Sie diesen Server installieren, müssen Sie einen Unsplash API-Zugriffsschlüssel (Access Key) erhalten:
Erstellen Sie ein Entwicklerkonto bei Unsplash
Registrieren Sie eine neue Anwendung
Holen Sie sich Ihren Zugriffsschlüssel von der Seite mit den Anwendungsdetails
Verwenden Sie diesen Schlüssel in den Konfigurationsschritten unten
Weitere Details finden Sie in der offiziellen Unsplash API-Dokumentation.
So installieren Sie den Unsplash Image Integration Server für Claude Desktop automatisch über Smithery:
Cursor IDE
npx -y @smithery/cli@latest install @hellokaton/unsplash-mcp-server --client cursor --key 7558c683-****-****Windsurf
npx -y @smithery/cli@latest install @hellokaton/unsplash-mcp-server --client windsurf --key 7558c683-****-****Cline
npx -y @smithery/cli@latest install @hellokaton/unsplash-mcp-server --client cline --key 7558c683-****-****# Clone the repository
git clone https://github.com/hellokaton/unsplash-mcp-server.git
# Navigate to project directory
cd unsplash-mcp-server
# Create virtual environment
uv venv
# Install dependencies
uv pip install .Cursor-Editor-Integration
Fügen Sie die folgende Konfiguration zur settings.json Ihres Cursor-Editors hinzu:
⚠️ Hinweis: Bitte passen Sie die folgende Konfiguration an Ihre tatsächliche Installation an:
Wenn uv nicht in Ihrem System-PATH enthalten ist, verwenden Sie einen absoluten Pfad (z. B. /path/to/uv)
./server.py sollte auf den tatsächlichen Speicherort Ihres Serverskripts angepasst werden (kann ein absoluter Pfad oder ein Pfad relativ zum Arbeitsbereich sein)
{
"mcpServers": {
"unsplash": {
"command": "uv",
"args": ["run", "--with", "fastmcp", "fastmcp", "run", "./server.py"],
"env": {
"UNSPLASH_ACCESS_KEY": "${YOUR_ACCESS_KEY}"
}
}
}
}{
"tool": "search_photos",
"query": "mountain",
"per_page": 5,
"orientation": "landscape"
}Golang: unsplash-mcp-server
Java: unsplash-mcp-server
Search for Unsplash photos
Args:
query: Search keyword
page: Page number (1-based)
per_page: Results per page (1-30)
order_by: Sort method (relevant or latest)
color: Color filter (black_and_white, black, white, yellow, orange, red, purple, magenta, green, teal, blue)
orientation: Orientation filter (landscape, portrait, squarish)
Returns:
List[UnsplashPhoto]: List of search results containing photo objects with the following properties:
- id: Unique identifier for the photo
- description: Optional text description of the photo
- urls: Dictionary of available image URLs in different sizes
- width: Original image width in pixels
- height: Original image height in pixels
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | ||
| page | No | ||
| per_page | No | ||
| order_by | No | relevant | |
| color | No | ||
| orientation | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden for behavioral disclosure. It mentions the tool searches photos but doesn't cover critical aspects like rate limits, authentication requirements, pagination behavior beyond parameters, or error conditions. The return format is described, but operational context is missing.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured and efficiently organized with clear sections for Args and Returns. Every sentence adds essential information without redundancy, making it easy to parse while maintaining completeness for the tool's functionality.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (6 parameters, 1 required) and lack of annotations/output schema, the description does well by detailing parameters and return format. However, it misses some behavioral context like rate limits or authentication needs, preventing a perfect score.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The description adds significant value beyond the input schema, which has 0% schema description coverage. It clearly explains each parameter's purpose, constraints (e.g., '1-30' for per_page), and valid values for enums like order_by, color, and orientation. This fully compensates for the schema's lack of descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Search for Unsplash photos'. It specifies the verb ('Search') and resource ('Unsplash photos'), making the function immediately understandable. However, without sibling tools to differentiate from, it cannot achieve the highest score of 5 which requires distinguishing from alternatives.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives, prerequisites, or constraints. It simply lists parameters and returns without context about appropriate use cases, making it minimally helpful for decision-making.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
search_photosScored across 1 tool
With only one tool, there is no possibility of confusion or overlap between tools. The single tool 'search_photos' has a clearly defined and distinct purpose.
The single tool follows a clear verb_noun pattern ('search_photos'), and with only one tool, there is perfect consistency by definition.
A single tool for an Unsplash server feels thin and incomplete for the domain. While search is a core function, typical photo services would include additional operations like getting photo details, downloading, or user-related actions.
The tool surface is severely incomplete for an Unsplash server. There are significant gaps: no ability to get photo details by ID, download photos, access user profiles, collections, or statistics. The server only supports search, which limits agent workflows.
Official ImageKit media library server: search, upload, organize, tag, share and manage assets.
Official ImageKit developer server: search the docs and build transformation URLs. No sign-in.
Search over 1.85 million captioned images using text, a reference image, or both. Upload a JPEG, PNG, or WebP image as base64, or provide a public HTTPS image URL. Find similar images from a Lightdrift asset ID and retrieve hosted file URLs, source licenses, and attribution. Connect with OAuth or an API key. Searches cost $0.005 ($5 per 1,000); image details are free. Setup: https://docs.lightdrift.ai/guides/images-mcp
The OpenRouter MCP server plugs OpenRouter into the AI tools you already use. Once connected, your assistant can pull live OpenRouter data (models, prices, your credits, rankings, and docs) and send quick test messages, all without leaving your editor.
