AIRA-SemanticScholar

This MCP server exposes a Semantic Scholar-powered suite of tools that lets AI agents perform literature search, author discovery, citation network analysis, and access/download full texts from supported sources. The server provides capabilities for comprehensive paper search (basic, advanced with filters, and batch retrieval), author profiling and publications, citation network traversal, field-specific browsing, and full-text access for arXiv and Wiley papers, along with DOI-based content resolution. Use the configured MCP endpoint to query these tools, enabling RAG-style tasks such as extracting key findings, tracing citation trajectories, and locating related works across disciplines. The included example configuration demonstrates how to run the server with Node.js, so clients can connect and invoke the semantic-scholar toolset from within an MCP-enabled workflow.

Prerequisites:

• Node.js installed on the host machine (recommended LTS)
• Access to Semantic Scholar API (optional but recommended for higher rate limits)
• Wiley Text and Data Mining (TDM) token if you plan to download/parse Wiley full-text PDFs
• Administrative access to your MCP environment (Smithery or equivalent) to deploy the server

Step 1: Install Node.js and dependencies

• Ensure Node.js is installed: node -v and npm -v
• Install any required dependencies for the MCP server setup (per your environment or Smithery instructions).

Step 2: Obtain and configure credentials

• If you plan to download Wiley full-texts, obtain a Wiley TDM Client Token and store it securely.
• If you have access to Semantic Scholar API, obtain an API key if required by your usage plan.

Step 3: Prepare the MCP server configuration

• Create or edit your MCP config (as shown in the example).
• Save the configuration in your MCP runner or Smithery deployment.

Step 4: Run the MCP server

• Use the command corresponding to your environment. For Node.js as shown:

# Example start (adjust path accordingly)
npx smithery start semantic-scholar-config.json

# Or directly run the Node script if your setup requires it:
node /path/to/build/index.js

Step 5: Verify and test

• Ensure the server appears in your MCP UI or registry as semantic-scholar.
• Run a basic query to search for papers or authors to verify connectivity with Semantic Scholar API.

Step 6: Optional integration with Wiley and DOIs

• If using Wiley full-text access, configure the WILEY_TDM_CLIENT_TOKEN in the env as shown in the mcp_config example.
• Test DOI resolution by requesting metadata/content via the DOI feature and validate the response.

Tips and considerations:

• Rate limits vary by source: Semantic Scholar API typically supports higher rates for authenticated requests; monitor usage to avoid throttling.
• For Wiley full-text access, ensure institutional access or appropriate subscriptions are in place; token-based access is required for full content retrieval.
• The mcp_config example uses a server key named semantic-scholar; you can rename it to match your deployment conventions.
• If you plan to scale, consider running multiple instances behind a load balancer and ensure your token/credentials are securely managed via environment variables.
• When testing, you can use the provided capabilities to perform batch paper retrievals (up to 500 papers per request) and analyze citation networks across multiple hops.