Semantic Scholar MCP Server

Instructions

Implementation Reference

• src/semantic_scholar_mcp/server.py:332-346 (handler)

  The core handler function that implements the semantic_scholar_search_authors tool. It queries the Semantic Scholar API for authors matching the input query, handles pagination, and formats the response in markdown or JSON.
  async def search_authors(params: AuthorSearchInput) -> str: """Search for academic authors by name.""" logger.info(f"Searching authors: {params.query}") response = await _make_request("GET", "author/search", params={"query": params.query, "offset": params.offset, "limit": params.limit, "fields": ",".join(AUTHOR_FIELDS)}) total, authors = response.get("total", 0), response.get("data", []) if params.response_format == ResponseFormat.JSON: return json.dumps({"query": params.query, "total": total, "authors": authors}, indent=2) lines = [f"## Author Search: \"{params.query}\"", f"**Found:** {total} authors", ""] for author in authors: lines.append(_format_author_markdown(author)) return "\n".join(lines)
• src/semantic_scholar_mcp/server.py:123-129 (schema)

  Pydantic input model defining the schema and validation for the tool's parameters.
  class AuthorSearchInput(BaseModel): model_config = ConfigDict(str_strip_whitespace=True, extra="forbid") query: str = Field(..., description="Author name to search", min_length=1, max_length=200) limit: int = Field(default=10, description="Max results", ge=1, le=100) offset: int = Field(default=0, description="Pagination offset", ge=0) response_format: ResponseFormat = Field(default=ResponseFormat.MARKDOWN, description="Output format")
• src/semantic_scholar_mcp/server.py:331-331 (registration)

  MCP decorator registering the search_authors function as the tool named semantic_scholar_search_authors.
  @mcp.tool(name="semantic_scholar_search_authors")
• src/semantic_scholar_mcp/server.py:251-267 (helper)

  Helper function to format author data into a readable markdown string, used in the tool's response.
  def _format_author_markdown(author: Dict[str, Any]) -> str: lines = [f"### {author.get('name', 'Unknown')}"] affiliations = author.get("affiliations") or [] if affiliations: lines.append(f"**Affiliations:** {', '.join(affiliations[:3])}") lines.append(f"**h-index:** {author.get('hIndex')} | **Papers:** {author.get('paperCount', 0)} | **Citations:** {author.get('citationCount', 0)}") if author.get("homepage"): lines.append(f"**Homepage:** {author['homepage']}") if author.get("url"): lines.append(f"**Profile:** [{author.get('authorId')}]({author['url']})") lines.append("") return "\n".join(lines)
• src/semantic_scholar_mcp/server.py:67-71 (helper)

  List of fields requested from the Semantic Scholar API for author data.
  AUTHOR_FIELDS: List[str] = [ "authorId", "externalIds", "url", "name", "aliases", "affiliations", "homepage", "paperCount", "citationCount", "hIndex" ]