Universal 1-Line Quickstart
Install and run find-research-papers-mcp across any modern agent runtime:
# ⚡ 1-Line Universal Installer (Auto-configures Claude Code, Cursor & Claude Desktop)
curl -fsSL https://papers.builditwithai.xyz/install | bash
# 🐍 Option 2: Run via Python (uvx)
uvx find-research-papers-mcp
# 📦 Option 3: Run via Node (npx)
npx -y find-research-papers-mcp
The 5 Scholarly Indexes
Instead of relying on LLM training weights that hallucinate citations, find-research-papers-mcp provides real-time verification across 250 million papers:
Instant access to latest preprints across CS, physics, mathematics, and quantitative biology with direct LaTeX and PDF extraction.
Biomedical and life sciences literature from the National Library of Medicine with structured MeSH terms and clinical trial links.
250M+ publications, author profiles, institutional affiliations, and concept taxonomies spanning all academic disciplines.
Authoritative DOI resolution, peer-reviewed journal metadata, funder registries, and publisher licensing data.
AI Agent Integration
Add papers to your agent environment:
Claude Code CLI
claude mcp add papers -- uvx find-research-papers-mcp
Cursor, Windsurf & Antigravity
Add to IDE settings:
{
"mcpServers": {
"papers": {
"command": "uvx",
"args": ["find-research-papers-mcp"]
}
}
}
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"papers": {
"command": "uvx",
"args": ["find-research-papers-mcp"]
}
}
}
Tool & Parameter Reference
| Tool Name | Parameters | Description |
|---|---|---|
search_papers |
query, sources, limit |
Unified multi-index search across arXiv, PubMed, OpenAlex, CrossRef, Semantic Scholar. |
get_paper |
doi or id |
Retrieves paper metadata, abstract, authors, and open-access PDF link. |
get_references |
doi |
Fetches complete bibliography and referenced papers via CrossRef. |
get_citations |
doi |
Fetches citing papers and citation graph via OpenAlex. |
verify_paper |
doi |
HEAD-checks landing page accessibility and cross-checks retraction databases. |
list_sources |
(none) | Returns live status and latency metrics for all 5 scholarly indexes. |
Citation Graph Traversal
AI models can traverse the forward and backward citation network in a single turn without manual re-prompting:
# Step 1: Discover foundational paper
papers.search_papers(query="Attention Is All You Need")
# Step 2: Traverse forward citations to find state-of-the-art descendants
papers.get_citations(doi="10.48550/arXiv.1706.03762", limit=5)