Scholar Search MCP

An MCP server for academic literature workflows in Claude, Cursor, and other MCP clients.

It combines Semantic Scholar + arXiv into one unified toolset, with fast parallel search, normalized outputs, source-aware deduplication, and practical research utilities (citations, references, author graph, recommendations, and arXiv source download).

Table of Contents

• Why this project
• What you get
• Demo videos
• Source strategy
• Install
• Quick setup (Claude Desktop)
• Quick setup (Cursor)
• Environment variables
• Tool list
• Testing with MCP Inspector
• Contributing
• License

Why this project

Most paper tools force you to choose one source or one API style. scholar-search-mcp focuses on a simple goal:

• One MCP server, multiple scholarly sources
• Free-first defaults (arXiv works without keys)
• LLM-friendly outputs for downstream reasoning and agent workflows
• Practical research actions, not only search

If you use AI agents for research, this gives you a cleaner and more reliable paper retrieval layer.

What you get

• Unified search (search_papers)
  • Queries Semantic Scholar and arXiv in parallel
  • Merges and deduplicates results by normalized title
  • Keeps richer metadata when overlaps occur
• Rich paper graph operations
  • Paper details, citations, references, author profile, author papers, recommendations
• Batch retrieval
  • Fetch up to 500 papers in one call (batch_get_papers)
• arXiv source workflow
  • Download and safely extract LaTeX/source tarballs via download_arxiv_source
• Built-in response caching
  • Improves repeated query latency and reduces API pressure
• Channel control by env vars
  • Turn Semantic Scholar / arXiv on or off without code changes

Demo videos

Agent writes a survey paper with Scholar Search MCP.

Source strategy

Current built-in sources:

• Semantic Scholar (metadata-rich, optional API key for better limits)
• arXiv (open and key-free)

Design principle:

1. Prefer open/public access paths first.
2. Support optional API keys when they improve stability or rate limits.
3. Keep outputs consistent for LLM consumption across different upstreams.

Install

pip install scholar-search-mcp

Requires Python 3.10+.

Quick setup (Claude Desktop)

Config file locations:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Minimal config:

{
  "mcpServers": {
    "scholar-search": {
      "command": "python",
      "args": ["-m", "scholar_search_mcp"],
      "env": {
        "SCHOLAR_SEARCH_ENABLE_SEMANTIC_SCHOLAR": "true",
        "SCHOLAR_SEARCH_ENABLE_ARXIV": "true"
      }
    }
  }
}

Recommended (with optional Semantic Scholar key):

{
  "mcpServers": {
    "scholar-search": {
      "command": "python",
      "args": ["-m", "scholar_search_mcp"],
      "env": {
        "SEMANTIC_SCHOLAR_API_KEY": "your-key",
        "SCHOLAR_SEARCH_ENABLE_SEMANTIC_SCHOLAR": "true",
        "SCHOLAR_SEARCH_ENABLE_ARXIV": "true"
      }
    }
  }
}

Quick setup (Cursor)

Add an MCP server in Cursor with the same:

• command: python
• args: ["-m", "scholar_search_mcp"]
• env: same variables as above

Environment variables

Example: run arXiv-only mode

{
  "SCHOLAR_SEARCH_ENABLE_SEMANTIC_SCHOLAR": "false",
  "SCHOLAR_SEARCH_ENABLE_ARXIV": "true"
}

Tool list

Testing with MCP Inspector

npm install -g @modelcontextprotocol/inspector
mcp-inspector python -m scholar_search_mcp

Contributing

Issues and pull requests are welcome.

If you want to contribute:

1. Fork the repo
2. Create a feature branch
3. Add tests or reproducible validation steps
4. Open a PR with clear before/after behavior

License

MIT

References

• Semantic Scholar API Docs
• arXiv API User Manual