MCP OCR Server

A production-grade OCR server built using MCP (Model Context Protocol) that provides OCR capabilities through a simple interface.

Features

Extract text from images using Tesseract OCR
Support for multiple input types:
- Local image files
- Image URLs
- Raw image bytes
Automatic Tesseract installation
Support for multiple languages
Production-ready error handling

# Using pip
pip install mcp-ocr

# Using uv
uv pip install mcp-ocr

Tesseract will be installed automatically on supported platforms:

python -m mcp_ocr

Configure Claude for Desktop: Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
    "mcpServers": {
        "ocr": {
            "command": "python",
            "args": ["-m", "mcp_ocr"]
        }
    }
}

Extract text from images:

# From file
perform_ocr("/path/to/image.jpg")

# From URL
perform_ocr("https://example.com/image.jpg")

# From bytes
perform_ocr(image_bytes)

List available OCR languages:

get_supported_languages()

git clone https://github.com/rjn32s/mcp-ocr.git
cd mcp-ocr

uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
uv pip install -e .

pytest

This project is licensed under the MIT License - see the LICENSE file for details.

Usage

As an MCP Server

Start the server:

python -m mcp_ocr

Configure Claude for Desktop: Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{ "mcpServers": { "ocr": { "command": "python", "args": ["-m", "mcp_ocr"] } } }

Available Tools

perform_ocr

Extract text from images:

# From file perform_ocr("/path/to/image.jpg") # From URL perform_ocr("https://example.com/image.jpg") # From bytes perform_ocr(image_bytes)

get_supported_languages

List available OCR languages:

get_supported_languages()