handwriting-ocr

The Handwriting OCR MCP Server acts as a bridge between MCP clients and the Handwriting OCR service. It accepts uploads of images or PDFs, monitors processing status, and returns the OCR results formatted as Markdown. You can interact with the server using the MCP client tools to upload documents, poll status, and retrieve the converted text. The server exposes a transcription workflow with a simple interface: upload a document, check status, then fetch the resulting text in Markdown.

Key capabilities include:

• Upload Document: Submit an image or PDF for OCR processing.
• Check Status: Retrieve the current processing state and any progress details.
• Get Text: Obtain the OCR result in Markdown format for easy integration into documents or notes.

To operate, ensure your MCP client points to the server instance configured under the handwriting-ocr path (as shown in the example configuration). You will need an API token for the Handwriting OCR Platform to authenticate requests. Once configured, you can automate uploads and result retrieval within your MCP workflows.

Prerequisites:

• Node.js 18.x or higher
• Access to an Handwriting OCR API token
• Basic MCP client/CLI familiarity

Installation via Smithery (recommended):

1. Install the MCP server via Smithery for Claude Desktop: npx -y @smithery/cli install @Handwriting-OCR/handwriting-ocr-mcp-server --client claude

Manual installation (Node.js):

1. Ensure you have Node.js installed (v18+).
2. Prepare an MCP configuration file (e.g., mcp_config.json) with the following content:

{
  "mcpServers": {
    "handwriting-ocr": {
      "command": "node",
      "args": [
        "/Users/your-user/path/to/handwriting-ocr/build/index.js"
      ],
      "env": {
        "API_TOKEN": "your-api-token"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

3. Start the MCP server using your environment (adjust paths as needed). Example:

node /Users/your-user/path/to/handwriting-ocr/build/index.js

4. Verify the server is running and accessible via the MCP client pointing to the configured server name (handwriting-ocr).

Environment variables:

• API_TOKEN: Your Handwriting OCR Platform API token, required for authentication.

Common issues:

• If the server fails to start, verify that the Node.js version is 18.x or higher and that the API_TOKEN is set in the environment.
• Ensure network access to the Handwriting OCR API endpoints from the host running the MCP server.

Configuration tips:

• Keep API_TOKEN secure; avoid committing it to version control.
• If deploying to production, consider rotating tokens and using a secret manager.
• You can extend the mcp_config to add additional servers or modify autoApprove behavior as needed.