Docling

An MCP server to help you "play with your documents" via Docling 🐥

python

18

GitHub Stars

python

Language

6 months ago

First Indexed

2 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": {
    "zanetworker-mcp-docling": {
      "command": "mcp-server-lls",
      "args": []
    }
  }
}

You have a document-processing MCP server that exposes a set of tools to convert, extract, and summarize documents. It runs locally or via a lightweight transport layer, and it integrates with Llama Stack to empower LLM applications with practical document understanding capabilities.

How to use

Start the server from your development environment and connect an MCP client to invoke its tools. Use the stdio transport for quick local runs, or run the server with SSE transport on a specific port if your client requires a networked endpoint.

How to install

Prerequisites: you need Python and pip installed on your system. You will also need a network-capable environment if you plan to expose the server over SSE.

Install the package in editable mode from your project directory.

pip install -e .

Run the server using the stdio transport (default). You can also run it via UV runtimes for different execution environments.

mcp-server-lls

If you prefer to run through UV, use the following command to start the server with stdio transport, or enable SSE on a port as shown.

uv run mcp-server-lls

uv run mcp-server-lls --transport sse --port 8000

Caching and notes

Processed documents are cached to improve performance for repeated requests. The cache is stored at ~/.cache/mcp-docling/ so you can inspect or manage it if needed.

Available tools

convert_document

Convert a document from a URL or local path to markdown format. Optional OCR can be enabled with enable_ocr and ocr_language specifies languages for OCR.

convert_document_with_images

Convert a document and extract embedded images. Optional OCR can be enabled with enable_ocr and ocr_language specifies languages for OCR.

extract_tables

Extract tables from a document as structured data.

convert_batch

Process multiple documents in batch mode. Provide a list of sources and optional OCR settings.

qna_from_document

Create a Q&A document from a URL or local path and format it as YAML. Requires IBM Watson X credentials set as environment variables: WATSONX_PROJECT_ID, WATSONX_APIKEY, WATSONX_URL.

get_system_info

Get information about the system configuration and acceleration status.