Local academic paper MCP server — 9-source search, multi-source download, AI analysis, translation, citation graph, code-based paper recommendation
{
"mcpServers": {
"scholar-mcp": {
"args": [
"scholar-mcp-server"
],
"command": "uvx"
}
}
}Are you the author?
Add this badge to your README to show your security score and help users find safe servers.
Local academic paper tool MCP server — 9-source search, multi-source download, AI-powered analysis, citation graph, code-based paper recommendation.
Is it safe?
No known CVEs for scholar-mcp-server.
No authentication — any process on your machine can connect.
MIT. View license →
Is it maintained?
Last commit 4 days ago. 33 stars.
Will it work with my client?
Transport: stdio. Works with Claude Desktop, Cursor, Claude Code, and most MCP clients.
Context cost
7 tools.
Local academic paper tool MCP server — 9-source search, multi-source download, AI-powered analysis, citation graph, code-based paper recommendation.
pip install scholar-mcp-server[all]
scholar-mcp-install --all
That's it. Restart your IDE and start using it.
| Tool | Description |
|---|---|
paper_search | 9-source concurrent search with relevance scoring (Semantic Scholar, OpenAlex, Crossref, PubMed, arXiv, CORE, Europe PMC, DOAJ, dblp) |
paper_download | Multi-source PDF download: Unpaywall → Publisher OA → arXiv → Sci-Hub → scidownl |
paper_batch_download | Batch download multiple papers by DOI list |
paper_ai_analyze | AI analysis — downloads PDF, extracts full text (up to 20 pages / 12k chars), sends to any OpenAI-compatible API |
paper_recommend | Scan your workspace code → multi-query auto-recommend related papers |
paper_citation_graph | Generate Mermaid citation/reference network visualization |
paper_health | Check download source availability |
Search results are ranked by a 4-factor composite score:
| Factor | Weight | Description |
|---|---|---|
| Query relevance | 0–40 | Title + abstract term matching |
| Citation impact | 0–30 | Log-scaled citation count |
| Source quality | 0–10 | Data source reliability weighting |
| Year recency | 0–15 | Boost for recent publications |
Deduplication uses DOI matching + Jaccard title similarity (≥0.7 threshold) across all 9 sources. Each source connector has built-in retry with exponential backoff.
paper_ai_analyze works with any OpenAI-compatible API. Set AI_API_BASE, AI_API_KEY, and AI_MODEL to point to your preferred provider.
git clone https://github.com/45645678a/Scholar-mcp.git
cd Scholar-mcp
pip install -r requirements.txt
python install.py --all
| Variable | Description | Required |
|---|---|---|
AI_API_KEY | API key for AI analysis | For paper_ai_analyze |
AI_API_BASE | API base URL (any OpenAI-compatible endpoint) | Optional (default: https://api.deepseek.com) |
AI_MODEL | Model name | Optional (default: deepseek-chat) |
UNPAYWALL_EMAIL | Email for Unpaywall API | Optional |
All free, no API keys required:
| Source | Coverage |
|---|---|
| Semantic Scholar | Broad academic (primary) |
| OpenAlex | 250M+ works, global |
| Crossref | DOI metadata |
| PubMed | Biomedical |
| arXiv | Physics, CS, Math |
| CORE | Open Access aggregator |
| Europe PMC | European biomedical |
| DOAJ | Open Access journals |
| dblp | Computer Science |
pip install .[all] pytest
pytest tests/ -v
40 tests covering search dedup, download chain, keyword extraction, and connector mocking.
This tool includes optional Sci-Hub integration for personal academic use. Sci-Hub may be illegal in some jurisdictions. Users are solely responsible for ensuring compliance with local laws and institutional policies. The authors do not endorse copyright infringement. If you are in a compliance-sensitive environment (university, company, lab), consult your institution's policy before using the Sci-Hub download source.
MIT
Run this in your terminal to verify the server starts. Then let us know if it worked — your result helps other developers.
uvx 'scholar-mcp-server' 2>&1 | head -1 && echo "✓ Server started successfully"
After testing, let us know if it worked:
paper_search9-source concurrent search with relevance scoring (Semantic Scholar, OpenAlex, Crossref, PubMed, arXiv, CORE, Europe PMC, DOAJ, dblp)
paper_downloadMulti-source PDF download: Unpaywall → Publisher OA → arXiv → Sci-Hub → scidownl
paper_batch_downloadBatch download multiple papers by DOI list
paper_ai_analyzeAI analysis — downloads PDF, extracts full text (up to 20 pages / 12k chars), sends to any OpenAI-compatible API
paper_recommendScan your workspace code → multi-query auto-recommend related papers
paper_citation_graphGenerate Mermaid citation/reference network visualization
paper_healthCheck download source availability
No known vulnerabilities.
Have you used this server?
Share your experience — it helps other developers decide.
Sign in to write a review.
A Model Context Protocol server for searching and analyzing arXiv papers
📖 MCP server for fetch deepwiki.com and get latest knowledge in Cursor and other Code Editors
Open source implementation and extension of Google Research’s PaperBanana for automated academic figures, diagrams, and research visuals, expanded to new domains like slide generation.
Offline access to TwitterAPI.io docs for AI assistants. 58 endpoints, 32 pages, 24 blog posts.
MCP Security Weekly
Get CVE alerts and security updates for Scholar Mcp and similar servers.
Start a conversation
Ask a question, share a tip, or report an issue.
Sign in to join the discussion.