Skip to main content
Glama
truaxki

Semantic Scholar MCP Server

by truaxki
OverviewSchemaRelated ServersScoreDiscussions
TypeScript
Remote

Semantic Scholar MCP Server

Remote MCP server with Streamable HTTP transport for Semantic Scholar API access.

Features

  • 🔌 Streamable HTTP Transport - Remote MCP server accessible anywhere

  • 💾 SQLite Caching - Persistent cache with TTL-based expiration

  • 🚦 Rate Limiting - Token bucket algorithm to prevent API quota exhaustion

  • 📊 Metrics - Prometheus-compatible metrics endpoint

  • 📝 Structured Logging - JSON logs with request IDs

  • 🔒 Security - API key via environment variable only (no CLI exposure)

Related MCP server: semantic-scholar-mcp

Quick Start

Installation

# Clone the repository
git clone https://github.com/your-org/semantic-scholar-mcp.git
cd semantic-scholar-mcp

# Install dependencies
npm install

# Build TypeScript
npm run build

Configuration

Set environment variables:

# Required: Semantic Scholar API key (get from https://www.semanticscholar.org/product/api)
export SEMANTIC_SCHOLAR_API_KEY="your-api-key"

# Optional: Server configuration
export PORT=3000
export HOST="0.0.0.0"
export LOG_LEVEL="info"

# Optional: Cache configuration
export CACHE_TTL_DAYS=7
export CACHE_MAX_SIZE=1073741824

# Optional: Rate limiting
export RATE_LIMIT_REQUESTS_PER_MINUTE=10
export RATE_LIMIT_BURST_SIZE=5

Running

# Production
npm start

# Development (with hot reload)
npm run dev

Authentication

Simple API Key (Default)

Uses environment variable for authentication. Good for personal use.

export SEMANTIC_SCHOLAR_API_KEY="your-api-key"

OAuth 2.1 (Enterprise)

For production multi-user deployments, use the OAuth 2.1 pattern:

Reusable OAuth Pattern: https://github.com/truaxki/mcp-oauth-pattern

The OAuth pattern provides:

  • OAuth 2.1 authentication with Supabase

  • Bearer token middleware

  • Consent page for user authorization

  • Per-user rate limiting (optional)

To add OAuth to this server:

  1. Clone https://github.com/truaxki/mcp-oauth-pattern

  2. Copy src/auth.ts, src/logger.ts from OAuth pattern

  3. Replace API key auth with Bearer token middleware

  4. Configure Supabase environment variables

Connect from Claude Code

Add to your .mcp.json:

{
  "mcpServers": {
    "semantic-scholar": {
      "type": "streamable-http",
      "url": "http://your-server:3000"
    }
  }
}

Endpoints

Endpoint

Description

POST /mcp

MCP protocol handler

GET /health

Health check

GET /metrics

Prometheus metrics

Available Tools

Tool

Description

search_papers

Search for papers with filters

get_paper

Get paper details by ID

get_authors

Get authors for a paper

get_citations

Get papers that cite a paper

get_references

Get papers referenced by a paper

batch_fetch

Fetch multiple papers efficiently

get_cache_stats

View cache statistics

clear_cache

Clear cached data

Deployment

Docker

# Build image
docker build -t semantic-scholar-mcp .

# Run container
docker run -d \
  --name semantic-scholar-mcp \
  -p 3000:3000 \
  -e SEMANTIC_SCHOLAR_API_KEY="your-key" \
  semantic-scholar-mcp

Railway Deployment

One-click deploy:

Deploy on Railway

Manual deployment:

  1. Connect repository

    • Go to Railway and sign in

    • Click "New Project" → "Deploy from GitHub repo"

    • Select truaxki/semantic-scholar-mcp

  2. Configure environment variables

    • In Railway dashboard, go to your service's "Variables" tab

    • Add:

      SEMANTIC_SCHOLAR_API_KEY=your-api-key-here
LOG_LEVEL=info
PORT=3000

  3. Deploy

    • Railway automatically detects Node.js and builds

    • Deploys with Streamable HTTP on port 3000

    • Gets free HTTPS endpoint automatically

  4. Connect from Claude Code

    {
  "mcpServers": {
    "semantic-scholar": {
      "type": "streamable-http",
      "url": "https://your-service.up.railway.app/mcp"
    }
  }
}

Railway pricing: Free tier available, $5/month for dedicated resources.

Manual Server (srv1338041)

# SSH to server
ssh user@srv1338041

# Clone and build
git clone your-repo
cd semantic-scholar-mcp
npm install
npm run build

# Create systemd service (see deploy/semantic-scholar-mcp.service)
sudo cp deploy/semantic-scholar-mcp.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable semantic-scholar-mcp
sudo systemctl start semantic-scholar-mcp

Architecture

┌─────────────────────────────────────────────────────┐
│              Semantic Scholar MCP Server             │
├─────────────────────────────────────────────────────┤
│                                                      │
│  ┌──────────────┐  ┌──────────────┐  ┌────────────┐ │
│  │   Express   │──▶│   MCP SDK   │──▶│  Tools    │ │
│  │   HTTP      │  │   Server    │  │  Registry │ │
│  └──────────────┘  └──────────────┘  └────────────┘ │
│         │                 │                           │
│         ▼                 ▼                           │
│  ┌──────────────┐  ┌──────────────┐                  │
│  │   Logging   │  │   Cache     │                  │
│  │  (Pino)     │  │ (SQLite)    │                  │
│  └──────────────┘  └──────────────┘                  │
│                                                      │
└─────────────────────────────────────────────────────┘
         │                 │                           │
         ▼                 ▼                           ▼
  HTTPS Endpoint    Persistent Cache    Rate Limited API

API Key

Get your free Semantic Scholar API key from: https://www.semanticscholar.org/product/api

Free tier: 100 requests/day

License

MIT

This server cannot be installed

F
license - not found
-
quality - not tested
D
maintenance

How are these scores calculated?

Maintenance

–Maintainers
–Response time
–Release cycle
–Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/truaxki/semantic-scholar-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server