OCR MCP Service

OCR MCP Service - Image text recognition for AI Agents

python

0

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "qiao-925-ocr-mcp-service": {
      "command": "ocr-mcp-server",
      "args": []
    }
  }
}

OCR MCP Service provides a unified interface to multiple OCR engines, enabling you to perform image text recognition through a single MCP endpoint. It streamlines setup, allows easy switching between engines, and lets Cursor coordinate and manage the OCR processes for you.

How to use

You will interact with this MCP service through your client, calling the available tools to recognize text in images. Start by ensuring the MCP server is available to your client and that you have at least one OCR engine installed. When you call a tool like recognize_image_paddleocr, the MCP server handles the underlying OCR processing and returns the recognized text. You can mix engines to compare results or use a single engine that best fits your language and accuracy needs. If you need general guidance, request the usage guide from within your client using the appropriate tool.

Using and testing the server

To test manually, you can start the server in a local stdio mode and verify it communicates with your client. Remember, the server is designed to be managed by the Cursor, so you usually don’t start it directly in production.

Getting more help

Access usage tips and templates by invoking the appropriate helper tools from your client, such as get_usage_guide or get_prompt_template, to customize prompts and workflows for OCR tasks.

Example prompts

Use the provided example prompt templates to structure image analysis tasks, and adapt them to your specific OCR workflow and target languages.

Available tools

recognize_image_paddleocr

Recognizes text in images using PaddleOCR, suitable for Chinese text and general usage in Chinese documents.

recognize_image_paddleocr_mcp

Uses paddleocr-mcp for MCP-based text recognition leveraging PaddleOCR under MCP management.

recognize_image_easyocr

Recognizes text in images using EasyOCR, supporting 80+ languages.

recognize_image_deepseek

Performs OCR with DeepSeek engine for higher accuracy on large or challenging documents.

get_prompt_template

Retrieves a generic prompt template for image analysis workflows to guide OCR and subsequent analysis.

get_usage_guide

Provides a comprehensive usage guide with tips and best practices for OCR tasks.