Paper Search · openags — Claude Code MCP server: what it does, how to install
Paper Search searches multiple open academic repositories, deduplicates metadata, and retrieves research papers. Useful when feeding verified scientific context into LLM workflows.
no license updated 2026-10-29
Open on GitHubWhat Paper Search does
Paper Search links MCP-compatible agents directly to academic literature repositories, including arXiv, PubMed, bioRxiv, Semantic Scholar, Crossref, and OpenAlex. The server deduplicates hits and outputs a normalized paper structure.
Its layout separates high-level tooling from individual platform connectors. The query engine handles multi-source concurrent calls, while the downloader follows an open-access fallback chain through CORE, OpenAIRE, Europe PMC, and Unpaywall. When speed matters more than broad coverage, search options -s fast and -s fastest narrow queries down to lower-latency indexes.
The server provides stdio, sse, and streamable-http transports. Running the package requires Python 3.10 or newer. Exposing the local HTTP transport without configuring OAuth creates an open network listener, which makes it insecure for public hosting.
What it helps with
- search academic papers across multiple public repositories concurrently
- download open-access research PDFs through sequential legal fallback links
- recover missing DOIs and bibliographic metadata via Crossref and Google Scholar
- inspect citation relationships and references using OpenAlex
Example prompts for the agent
- Find recent papers on quantum error correction across arXiv and Semantic Scholar.
- Download the open-access PDF for DOI 10.1038/s41586-020-2649-2 and extract its key takeaways.
- Pull the top citing works for this paper from OpenAlex with a bounded budget.
How to install Paper Search
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Paper Search
How to run Paper Search with Claude Code?
The server connects to MCP clients using stdio, sse, or streamable-http transports. It requires Python 3.10 or higher and can be launched via Smithery, uvx, uv, pip, npx, or Docker.
What does Paper Search do?
It queries multiple open research repositories, deduplicates entries, and fetches available full-text papers into a uniform data structure. Supported platforms include arXiv, PubMed, bioRxiv, Semantic Scholar, Crossref, and OpenAlex.
Are paid API keys required to use Paper Search?
No, the server prioritizes public, free data sources and functions without mandatory subscriptions. Optional API keys for services like Semantic Scholar only help improve rate limits and metadata quality.
Similar MCP servers
All MCP servers for Science and research →Paper Search by other authors comes first.