home / mcp / dblp mcp server

DBLP MCP Server

A Model Context Protocol (MCP) server that provides access to the DBLP computer science bibliography database for Large Language Models.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "szeider-mcp-dblp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/mcp-dblp/",
        "run",
        "mcp-dblp"
      ]
    }
  }
}

You can access the DBLP computer science bibliography through an MCP server that integrates DBLP data with model context workflows. This server lets you search publications, fetch BibTeX entries, perform fuzzy matches, analyze statistics, and export BibTeX collections to a file for precise bibliographic handling. It is designed to work smoothly with MCP clients so you can incorporate DBLP data directly into your AI workflows.

How to use

Use an MCP client to interact with the mcp_dblp server. You can search publications with boolean queries, perform fuzzy title or author name matching, retrieve author publication lists, fetch venue information, calculate statistics from results, and manage a BibTeX collection that can be exported to a .bib file. All BibTeX entries can be fetched directly from DBLP for maximum accuracy, and you can export the complete collection when you are ready to cite sources.

How to install

Prerequisites: ensure you have Python 3.11+ and a runtime for executing MCP servers (the MCP flow in this setup uses the uv runner). You will also need Git to clone the source and a shell to run commands.

# Prerequisites
python3.11 --version
git --version

# Clone the MCP-DBLP project
git clone https://github.com/szeider/mcp-dblp.git
cd mcp-dblp

# Create and activate a Python virtual environment (development steps)
uv venv
source .venv/bin/activate

# Install the MCP package in editable mode
uv pip install -e .

# Configure Claude Desktop (example). Place this JSON in your Claude Desktop config location.

Configuration and usage notes

To run the local MCP server in a development environment, use the runtime command and arguments shown for the stdio configuration. This starts the mcp-dblp server from your local directory and runs it under the uv ecosystem.

{
  "mcpServers": {
    "mcp_dblp": {
      "type": "stdio",
      "name": "mcp_dblp",
      "command": "uv",
      "args": ["--directory", "/path/to/mcp-dblp/", "run", "mcp-dblp"]
    }
  }
}

Using BibTeX export and libraries

The server supports a BibTeX export workflow. Add BibTeX entries by DBLP key, then export all collected entries to a .bib file. The export writes a BibTeX file to a path you specify and clears the collection after a successful export.

Key actions you can perform: - Add a BibTeX entry using a DBLP key - Export the collected BibTeX entries to a file with an absolute path - Retrieve BibTeX entries directly from DBLP with a timeout safeguard for reliability.

Troubleshooting and tips

If a run times out while fetching BibTeX entries, you’ll receive an error and the entry will not be added to the collection. Ensure the local path provided for exports exists or can be created, and verify that the mcp-dblp directory is accessible by the runtime. If you change the directory path, update the stdio configuration accordingly.

Available tools

get_instructions

Get usage instructions and workflow guidance for using the MCP server.

search

Query DBLP publications using boolean strings with optional year and venue filtering.

fuzzy_title_search

Find publications by fuzzy matching on titles with a similarity threshold.

get_author_publications

Retrieve publications for a specific author with optional similarity and count control.

get_venue_info

Fetch detailed information about a publication venue.

calculate_statistics

Compute statistics from a list of publication results.

add_bibtex_entry

Add a BibTeX entry to the collection by a DBLP key and customize the citation key.

export_bibtex

Export all collected BibTeX entries to a .bib file at a specified path.