We are currently generating PDF files from DOCX templates using both JSReport and a custom Node.js script that utilizes LibreOffice (via libre.convert) and unoconv. Below are the two approaches we've tested:

Approach 1: JSReport Configuration

{
  "template": {
    "recipe": "docx",
    "engine": "handlebars",
    "unoconv": {
      "format": "pdf"
    },
    "docx": {
      "templateAsset": {
        "content": "<base64-encoded-wordDocumentBuffer>"
      }
    }
  },
  "data": "Objects"
}

Approach 2: Node.js Script Using LibreOffice (LibreOffice Convert)

function convertDocxToPdf(docxBuffer) {
  return new Promise((resolve, reject) => {
    const extend = ".pdf";
    libre.convert(docxBuffer, extend, undefined, (err, pdfBuffer) => {
      if (err) {
        return reject(
          new Error("Error during DOCX to PDF conversion: " + err?.message)
        );
      }
      resolve(pdfBuffer);
    });
  });
}

Issue Observed:

On Linux systems, the conversion is fast and works as expected.
On Windows systems with high configuration (e.g., 16GB+ RAM, SSD), the performance is acceptable.
However, on Windows systems with lower specs (e.g., 8GB RAM), both methods take around 15–20 seconds for a single DOCX to PDF conversion.

We have tested with:

LibreOffice direct CLI conversion
Unoconv package
JSReport with unoconv recipe

But in all cases, the conversion time is consistently slow on low-end Windows systems.
Here, We are looking to consistently reduce conversion time to under 10 seconds on all environments.