Clarifai MCP Server Local

🚀 Clarifai MCP Server Project Overview

This project is a Clarifai MCP server that provides a series of interfaces for image generation and search, enabling users to generate images based on text prompts and search for images on the Clarifai platform using local file paths or network URLs.

✨ Features

Tool Interfaces

generate_image

• Function Description: Generate an image using the given text.
• Input Parameters:
  • prompt (string): Text prompt, e.g., "a beautiful sunset over the mountains".
  • width (int, optional): Image width, default is 512.
  • height (int, optional): Image height, default is 512.
  • quality (string, optional): Image quality, available values include "standard", "high", "ultra".
• Output Result:
  • Returns the file path of the generated image.

clarifai_image_by_path

• Function Description: Search for images on the Clarifai platform based on the local file path.
• Input Parameters:
  • filepath (string): Local path of the image file.
  • width (int, optional): Width requirement for the search results.
  • height (int, optional): Height requirement for the search results.
• Output Result:
  • Returns a list of matching images.

clarifai_image_by_url

• Function Description: Search for images on the Clarifai platform based on the network URL address.
• Input Parameters:
  • url (string): Network address containing the image.
  • width (int, optional): Width requirement for the search results.
  • height (int, optional): Height requirement for the search results.
• Output Result:
  • Returns a list of matching images.

Resource Interfaces

Resource List Request (resources/list)

• Supports specifying different search types through URI parameters:
  • /search/text: Text content search.
  • /search/url: Network link search.
• Supports specifying filtering conditions through query parameters, e.g., width, height, format.

Resource Read Request (resources/read)

• Returns specific resource information based on the given resource ID.

🔧 Technical Details

Project Architecture Diagram

cmd/server: Entry point for starting the server, initializing configurations, and registering handlers.
config: Handles command-line parameters and stores configurations such as PAT and output path.
mcp: Responsible for the communication logic of the MCP protocol, processing JSON-RPC requests on standard input/output.
clarifai: Encapsulates the gRPC client logic for calling the Clarifai API and includes authentication functionality (adding PAT to the request header).
tools: Implements various MCP tool handlers (e.g., generate_image, clarifai_image_by_path), using the clarifai client to perform specific operations.
resources: Handles MCP resource-related requests (e.g., resources/list and resources/read), calling the Clarifai API to obtain data.
utils: Provides auxiliary functions, such as file system operations.

Request Flow Diagram

1. The client sends a JSON-RPC request to the server.
2. The server parses the request and distributes it to different handlers based on the type.
3. The handler calls the Clarifai API to perform specific operations.
4. The result is returned to the client.

System Architecture and Component Relationships

cmd/server -> Responsible for starting and initializing configurations.
config -> Stores runtime configurations.
mcp -> Communication layer, handling protocol details.
clarifaiclient -> Encapsulates API call logic.
tools -> Handles specific tool requests (e.g., image generation).
resources -> Handles resource-related requests (e.g., search).
utils -> Auxiliary functions (e.g., file operations).

Through the above modules and interfaces, the system implements a complete set of functions from image generation to content retrieval and provides services externally through the MCP protocol.