Mcp Scholaris

io.github.penny4nonsensePublished by GitHub account penny4nonsensev0.1.4Updated Oct 24, 2026

Academic paper search and retrieval via arXiv, Semantic Scholar, PubMed, Unpaywall.

Overview

AI-generated overview

Local MCP server that lets an assistant search arXiv, Semantic Scholar and PubMed and fetch open-access full texts by arXiv ID, PMID or DOI.

What you can do
  • Search papers by topic, author or keyword across arXiv, Semantic Scholar and PubMed, choosing sources and result counts.
  • Fetch full text of a paper by arXiv ID, PubMed ID or DOI, trying arXiv, PubMed Central and Unpaywall in order.
  • Retrieve open-access PDFs and PMC article text rather than paywalled content.
When to use it
Useful when you want an assistant to look up academic literature, find papers on a topic, or pull open-access full texts while researching, without leaving the chat.
Requirements
Runs locally over stdio; install the mcp-scholaris Python package (pip) and register the scholaris command in the MCP client. Network access to arXiv, Semantic Scholar, PubMed and Unpaywall is needed. An optional SEMANTIC_SCHOLAR_API_KEY in a .env file raises Semantic Scholar rate limits; the key is free.
Before you install
Without an API key, Semantic Scholar requests are rate-limited and may fail or be slow. Queries and identifiers are sent to the third-party services arXiv, Semantic Scholar, PubMed and Unpaywall. The optional SEMANTIC_SCHOLAR_API_KEY is read from a .env file in the working directory, so keep that file out of version control.

Installation

In SourceWeft

  1. Open Mcp Scholaris in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

mcp-scholaris

An MCP server for retrieving academic papers from open access sources. Provides search across arXiv, Semantic Scholar, and PubMed, with full text retrieval via arXiv PDFs, PubMed Central (BioC API), and Unpaywall. No paywalls, no grey area — legitimate open access only.

Installation

bash
pip install mcp-scholaris

Tools

search_papers

Search for academic papers by topic, author, or keyword.

Parameters:

  • query (required) — search terms
  • sources (optional) — list of sources to search: arxiv, semantic_scholar, pubmed. Defaults to all three.
  • max_results (optional) — maximum results per source. Defaults to 5.

fetch_paper

Fetch the full text of a paper. Tries open access sources in order: arXiv → PubMed Central → Unpaywall. Provide at least one identifier.

Parameters:

  • arxiv_id — e.g. 2301.00001 or arxiv:2301.00001
  • pubmed_id — PubMed ID (PMID), e.g. 36383508
  • doi — e.g. 10.1371/journal.pone.0276755

Configuration

Semantic Scholar works without an API key but is rate-limited. For higher limits, create a .env file in your working directory:

SEMANTIC_SCHOLAR_API_KEY=your_key_here

API keys are free at semanticscholar.org.

Usage with an MCP client

Add to your MCP client configuration:

json
{  "mcpServers": {    "scholaris": {      "command": "scholaris"    }  }}

Or run directly:

bash
scholaris

The server communicates over stdio using the MCP protocol (JSON-RPC 2.0).

Sources

SourceSearchFull Text
arXiv✓✓ PDF
Semantic Scholar✓✓ when OA PDF available
PubMed✓✓ via BioC API (PMC articles)
Unpaywall—✓ for any DOI with OA version

License

MIT

Source: README.md at commit 0a3e141

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.1.4LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts