TextIn OCR MCP

What is TextIn OCR MCP about?

TextIn OCR MCP provides three core tools for document handling: (1) recognition_text extracts raw text from images, Word files, and PDFs; (2) doc_to_markdown converts those files into Markdown format; (3) general_information_extration identifies and returns user‑specified key information and table data as JSON.

How to use TextIn OCR MCP?

1. Register on the TextIn platform and obtain an APP_ID and APP_SECRET.
2. Start the MCP server via npx with the configuration shown in the serverConfig section.
3. Call the desired tool via the MCP API, providing a path that can be a local file path or an HTTP/HTTPS URL (protected URLs are not supported).
4. Receive the result – plain text, Markdown, or a JSON object containing extracted fields.

Key features of TextIn OCR MCP

• Multi‑format support: PDF, Word, Excel, and common image types (JPG, JPEG, PNG, BMP).
• Three specialized endpoints: plain text OCR, document‑to‑Markdown conversion, and flexible information extraction.
• Configurable extraction: specify key and table_header arrays to target particular data.
• Simple deployment: run with a single npx command; environment variables control authentication and timeout.
• MIT licensed: free to modify and redistribute.

Use cases of TextIn OCR MCP

• Automating receipt, invoice, or ID scanning pipelines.
• Converting legacy PDFs and scanned documents into editable Markdown for knowledge bases.
• Pulling structured data (e.g., contract dates, amounts, tables) from bulk document stores.
• Building search indexes by first extracting plain text from heterogeneous file types.

FAQ from the TextIn OCR MCP

Q: Can I process password‑protected PDFs? A: No. The service only accepts files that are directly readable; protected resources are not supported.

Q: What is the maximum request timeout? A: The default timeout is set to 600 seconds (MCP_SERVER_REQUEST_TIMEOUT=600000 ms), but it can be adjusted via the environment variable.

Q: Do I need to host any additional services? A: No. All functionality is bundled in the npm package; just run the server with the provided npx command.

TextIn OCR MCP

English | 中文

TextIn OCR MCP Server

TextIn MCP Server is a tool for extracting text and performing OCR on documents, including document text recognition, ID recognition, and invoice recognition. It also supports converting documents into Markdown format.

Tools

• recognition_text

  • Text recognition from images, Word documents, and PDF files.
  • Inputs:
    • path (string, required): file path or a URL (HTTP/HTTPS) pointing to a document
  • Return: Text of the document.
  • Supports conversion for:
    • PDF
    • Image (Jpeg, Jpg, Png, Bmp)

• doc_to_markdown

  • Convert images, PDFs, and Word documents to Markdown.
  • Inputs:
    • path (string, required): file path or a URL (HTTP/HTTPS) pointing to a document
  • Return: Markdown of the document.
  • Supports conversion for:
    • PDF
    • Microsoft Office Documents (Word, Excel)
    • Image (Jpeg, Jpg, Png, Bmp)

• general_information_extration

  • Automatically identify and extract information from documents, or identify and extract user-specified information.
  • Inputs:
    • path (string, required): file path or a URL (HTTP/HTTPS) pointing to a document
    • key (string[], optional): The non-tabular text information that the user wants to identify, input format is an array of strings.
    • table_header (string[], optional): The table information that the user wants to identify, input format is an array of strings.
  • Return: The key information JSON.
  • Supports conversion for:
    • PDF
    • Microsoft Office Documents (Word, Excel)
    • Image (Jpeg, Jpg, Png, Bmp)

When the input is a URL, it does not support handling access to protected resources.

Setup

APP_ID and APP_SECRET

Click here to register for a TextIn account.

Get Textin APP_ID and APP_SECRET by following the instructions here.

NPX

{
  "mcpServers": {
    "textin-ocr": {
      "command": "npx",
      "args": [
        "-y",
        "@intsig/server-textin"
      ],
      "env": {
        "APP_ID": "<YOUR_APP_ID>",
        "APP_SECRET": "<YOUR_APP_SECRET>",
        "MCP_SERVER_REQUEST_TIMEOUT": "600000"
      },
      "timeout": 600
    }
  }
}

License

This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.