nutrient-document-engine

The Nutrient Document Engine MCP Server exposes document processing capabilities via natural language, enabling interactions with document collections through AI agents or direct human use. It provides tools for content access (text extraction, search, page rendering, form data and table extraction), document editing (splitting documents, adding watermarks, duplicating documents, filling PDFs), security (redactions with preset patterns or custom regex), and annotations (highlights, notes, stamps). You can query documents, perform extraction and transformation tasks, and drive workflows from Claude Desktop, LangGraph agents, or OpenAI-compatible interfaces using the MCP protocol. To get started, configure the MCP server in Claude Desktop or your preferred MCP-enabled platform and point it at the node running the server. Once connected, you can ask questions like “List my documents and extract text from the contract” or “Redact SSNs from these forms,” and the MCP server will orchestrate the document engine to perform the requested actions.

Typical workflows include uploading documents to the Document Engine dashboard, then using natural language prompts to extract data, fill forms, redact sensitive content, or generate annotated outputs. The server is designed to work with multiple AI platforms through the MCP interface, enabling cross-tool collaboration for complex document processing tasks. It also exposes configuration options and environment variables to tailor authentication, base URLs, and licensing as you integrate with Claude Desktop or other agents.

Prerequisites:

• Docker and Docker Compose installed on your machine
• Node.js 18+ (for local MCP client usage with npx)
• Access to CLAUDE Desktop or another MCP-compatible interface

1. Clone the repository and start the Document Engine locally

• Prerequisites assume you’re using the Nutrient Document Engine with Docker Compose as shown in the README.

git clone https://github.com/PSPDFKit/nutrient-document-engine-mcp-server.git
cd nutrient-document-engine-mcp-server
docker-compose up -d

2. Install and configure the MCP client (Claude Desktop example)

• Install Claude Desktop per the instructions and ensure Docker-based services are up.
• In Claude Desktop, add a new MCP server configuration using npx to load the MCP server package:

"mcpServers": {
  "nutrient-document-engine": {
    "command": "npx",
    "args": ["-y", "@nutrient-sdk/document-engine-mcp-server"],
    "env": {
      "DASHBOARD_USERNAME": "your-dashboard-username",
      "DASHBOARD_PASSWORD": "your-dashboard-password",
      "DOCUMENT_ENGINE_BASE_URL": "http://localhost:5000",
      "DOCUMENT_ENGINE_API_AUTH_TOKEN": "your-api-token"
    }
  }
}

3. Verify the server is reachable

• Open the Document Engine dashboard at http://localhost:5100/dashboard to upload and manage documents.
• Ensure Claude Desktop (or your MCP client) is connected to the nutrient-document-engine MCP server via the configuration above.

4. Optional licensing configuration

• If you need to remove evaluation limitations, follow the Nutrient licensing guide and set ACTIVATION_KEY as an environment variable in your deployment.

Notes:

• Replace placeholder values with your actual credentials, base URL, and tokens.
• If you’re deploying in production, consider using a dedicated server, TLS termination, and secure storage for credentials.

Tips and common considerations:

• The MCP server uses environment variables to control access and to configure the base URL of the Document Engine. Ensure DOCUMENT_ENGINE_BASE_URL points to your running engine instance.
• For Claude Desktop, ensure the config JSON is correctly placed under Settings → Developer → Edit Config and that the MCP server is accessible from the Claude environment.
• If you run into authentication issues, verify that DASHBOARD_USERNAME and DASHBOARD_PASSWORD match the credentials configured for the Nutrient Document Engine dashboard.
• The MCP server supports a variety of tools (text extraction, form data extraction, redactions, annotations, splitting/merging, etc.). You can customize capabilities by enabling features in the underlying Document Engine and adjusting licensing.
• When deploying in a cloud or containerized environment, consider mounting a persistent volume for documents and ensuring proper network access between the MCP server and the Document Engine service.
• Review the documentation references inside the repository for advanced configuration (e.g., environment variables, transport modes, and dashboard settings).