home / mcp / daniel lightrag mcp server

Daniel LightRAG MCP Server

A comprehensive MCP server for LightRAG integration with 22 tools for document management, querying, knowledge graph operations, and system management

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "desimpkins-daniel-lightrag-mcp": {
      "command": "python",
      "args": [
        "-m",
        "daniel_lightrag_mcp"
      ],
      "env": {
        "LOG_LEVEL": "INFO",
        "LIGHTRAG_API_KEY": "YOUR_API_KEY",
        "LIGHTRAG_TIMEOUT": "30",
        "LIGHTRAG_BASE_URL": "http://localhost:9621"
      }
    }
  }
}

You operate the Daniel LightRAG MCP Server, which bridges LightRAG with your MCP client to provide 22 tools across document management, querying, knowledge graph operations, and system management. This server lets you insert and retrieve documents, run advanced queries, manage knowledge graph entities and relations, and monitor health and caches, all through a consistent MCP interface.

How to use

You connect your MCP client to the Daniel LightRAG MCP Server and use the available tools to manage documents, run queries, and oversee knowledge graph data. Start by ensuring LightRAG is running on http://localhost:9621, then launch the MCP server configured to invoke the Python module that exposes the MCP interface. Use the MCP client to call tools by name and provide the required parameters. If you need to verify connectivity, run the health check tool and confirm a healthy response.

How to install

# Basic installation
pip install -e .

# With development dependencies
pip install -e ".[dev]"

Configuration and usage notes

The MCP server is started as a local stdio server that runs the Python module for the MCP interface. You configure your MCP client to connect to this local runtime and pass environment settings to align with your LightRAG instance. The following configuration example shows how to register the server in an MCP client, using Python as the launcher and the module name daniel_lightrag_mcp.

{
  "mcpServers": {
    "daniel_lightrag": {
      "command": "python",
      "args": ["-m", "daniel_lightrag_mcp"]
    }
  }
}

Security and environment setup

The server relies on LightRAG being accessible at the address you configure. You can provide an API key and base URL to regulate access and timeouts. In typical setups you pass environment variables to the MCP process to point it at LightRAG and to supply optional security credentials.

Troubleshooting quick checks

If you run into issues, verify LightRAG is reachable, then test the MCP server start sequence. Check that the environment variables are set correctly and that the server process starts without errors. If a specific tool reports an error, review the error details to determine whether it is a connectivity, authentication, or parameter validation issue.

Example workflows

You can perform end-to-end workflows that involve inserting documents, querying for insights, and examining the knowledge graph to confirm relationships between entities. Start with a health check, add documents with the appropriate insert tool, then run a query and explore the knowledge graph structure to verify results.

Available tools

insert_text

Insert a single text document into LightRAG with the provided content.

insert_texts

Insert multiple text documents into LightRAG, including optional titles and metadata.

upload_document

Upload a document file to LightRAG from a given file path.

scan_documents

Scan LightRAG for newly added documents to update the index.

get_documents

Retrieve all documents stored in LightRAG.

get_documents_paginated

Retrieve documents with pagination support for page number and page size.

delete_document

Delete a specific document by its ID from LightRAG.

clear_documents

Remove all documents from LightRAG.

query_text

Execute a text-based query against LightRAG with optional modes and context behavior.

query_text_stream

Stream results for a text query from LightRAG.

get_knowledge_graph

Retrieve the full knowledge graph from LightRAG.

get_graph_labels

Fetch all labels available in the knowledge graph.

check_entity_exists

Check whether a named entity exists in the knowledge graph.

update_entity

Update properties for an existing knowledge graph entity.

update_relation

Update properties for an existing knowledge graph relation.

delete_entity

Delete an entity from the knowledge graph.

delete_relation

Delete a relation from the knowledge graph.

get_pipeline_status

Obtain current pipeline status from LightRAG.

get_track_status

Retrieve status information for a specific tracked operation.

get_document_status_counts

Get counts of documents by processing status.

clear_cache

Clear the LightRAG cache to reset in-memory indices.

get_health

Check the LightRAG server health and responsiveness.