Paper Search MCP

Instructions

Implementation Reference

• paper_search_mcp/server.py:1253-1264 (handler)

  The 'download_ssrn' handler in 'server.py', which acts as the wrapper function for the tool. Note that it delegates to the 'SSRNSearcher' and documentation notes it is a metadata-only connector.

  async def download_ssrn(paper_id: str, save_path: str = "./downloads") -> str:
      """Download PDF for a paper from SSRN.
  
      Note: SSRN connector is metadata-only and download is not supported.
  
      Args:
          paper_id: SSRN paper identifier.
          save_path: Directory to save the PDF (unused).
      Returns:
          str: Error message from metadata-only SSRN connector.
      """
      return ssrn_searcher.download_pdf(paper_id, save_path)

• paper_search_mcp/academic_platforms/ssrn.py:106-145 (handler)

  The 'SSRNSearcher.download_pdf' implementation in 'ssrn.py'. This contains the logic (or lack thereof, as it is a best-effort/metadata-only service) for downloading PDFs from SSRN.

  def download_pdf(self, paper_id: str, save_path: str = "./downloads") -> str:
      """Download PDF for an SSRN paper when a public direct link is available.
  
      SSRN frequently requires login for PDF delivery. This method is
      best-effort only and returns an explanatory message when no accessible
      public PDF link can be resolved.
  
      Args:
          paper_id: SSRN ID in ``ssrn:<id>`` format, raw numeric id, or URL.
          save_path: Directory to save downloaded PDF.
  
      Returns:
          Saved PDF path on success, otherwise an explanatory message.
      """
      abstract_id = self._extract_abstract_id(paper_id)
      if not abstract_id:
          return f"Invalid SSRN paper id: {paper_id}"
  
      pdf_url = self._resolve_pdf_url(abstract_id)
      if not pdf_url:
          return (
              f"No publicly accessible SSRN PDF URL found for {abstract_id}. "
              "The paper may require SSRN login or restricted access."
          )
  
      os.makedirs(save_path, exist_ok=True)
      output_path = os.path.join(save_path, f"ssrn_{abstract_id}.pdf")
  
      try:
          response = self.session.get(pdf_url, stream=True, timeout=60)
          response.raise_for_status()
  
          content_type = (response.headers.get("content-type") or "").lower()
          first_chunk = next(response.iter_content(chunk_size=1024), b"")
          if "pdf" not in content_type and not first_chunk.startswith(b"%PDF"):
              return (
                  f"Resolved SSRN URL is not a direct PDF ({pdf_url}). "
                  "This likely requires browser login."
              )