# scopus-mcp [Health: Active]

**Category:** 🔬 Research  
**Repository:** https://github.com/qwe4559999/scopus-mcp  
**GitHub Stars:** 47  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/scopus-mcp

## Description
A Model Context Protocol server for the Elsevier Scopus API

## Tools
Capabilities this server exposes over MCP:

- **query**
- **count**
- **sort**
- **coverDate**
- **scopus_id**
- **author_id**

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "scopus-mcp": {
    "command": "uvx",
    "args": ["scopus-mcp"],
    "env": {
      "SCOPUS_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `SCOPUS_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What scopus-mcp MCP server does

The scopus-mcp MCP server exposes selected Elsevier Scopus operations to an MCP-compatible assistant. It can search the Scopus database using standard query syntax, return a chosen number of results, and apply a sort order such as publication date. For example, a query can target titles with syntax such as `TITLE("Artificial Intelligence")`.

It also supports document and researcher lookups. Given a Scopus document identifier, the server retrieves detailed information including abstract data. Given a Scopus Author ID, it retrieves the corresponding author profile. These capabilities suit literature discovery, paper review, and basic author-focused research tasks.

## How it works

The server acts as a local MCP process and sends requests to the Elsevier Scopus API using an API key. An MCP client presents the available operations to an AI assistant, which can then supply search terms, pagination-related result counts, sort options, document IDs, or author IDs as appropriate.

The repository describes three higher-level operations: Scopus search, abstract and document detail retrieval, and author profile retrieval. The exposed parameters include `query`, `count`, `sort`, `scopus_id`, and `author_id`. Search results default to five items when no result count is supplied. The client configuration starts the process with `uvx scopus-mcp`, although the module can also be launched directly with Python.

## Setup and configuration

Using the scopus-mcp MCP server requires an Elsevier Scopus API key. The README directs users to the Elsevier Developer Portal to request one and notes that access commonly depends on an institutional or organizational Elsevier subscription. Public email domains may be rejected during the key request process.

For a local project setup, create `config.json` in the repository root, or copy the example file, and add the key under `api_key`. You then adjust the path in `MCP_tool_config.json` and import that configuration into the MCP client. The quick-start configuration instead passes the key as the `SCOPUS_API_KEY` environment variable.

The recommended launch method uses `uvx`, so `uv` must be installed for that route. Alternatively, install Python 3.10 or newer, run `pip install .`, and start the module with `python -m scopus_mcp.server`. Claude Desktop configuration files differ by operating system; Cursor can add the server as a command-type, stdio MCP server. The README also provides configuration instructions for Trae.

## Tools and capabilities

The scopus-mcp MCP server supports these functional areas:

- Search Scopus with a query string and optional result count and sort order.
- Retrieve detailed information for a document using its Scopus ID.
- Retrieve profile information for an author using a Scopus Author ID.
- Use `coverDate` as a documented sort option for search results.

The server is intended for retrieving information available through the Scopus API. It does not document facilities for modifying Scopus records, managing subscriptions, or performing general web searches.

## Limitations and notes

API availability depends on Elsevier credentials and the access granted to those credentials. An MCP client alone does not provide Scopus access. The documented setup requires a valid API key, and the README warns that institutional access may be relevant.

The repository lists Python 3.10 or newer for manual installation. The zero-setup path is not literally dependency-free because it relies on the `uv` package manager. Configuration paths and slash conventions must be adjusted for the host operating system.

_Full upstream README: https://allmcps.com/mcp/scopus-mcp/readme_