Troubleshooting desk
Troubleshoot Markdown document conversion
A practical checklist for empty previews, failed jobs, ZIP files, fonts, tables, formulas, diagrams, downloads, and output differences.
Check in order
Find the first failing stage, then fix only that stage
A failed conversion is easier to diagnose when the source, preview, request, job, and download are checked separately.
01The preview is empty when the page first opens
Wait for the initial sample to finish loading, then confirm the source textarea contains Markdown. If editing the source immediately makes the preview appear, check whether the workspace script or API configuration loaded before testing again.
02The request fails before a job is created
Check that the source is not empty, a Markdown entry has been selected from a ZIP, and the chosen output is supported by the endpoint. Do not retry the same oversized or unsafe package unchanged.
03The preview and downloaded file look different
The HTML workspace downloads the same preview snapshot and collected styles. DOCX and PDF use their own native writers, so recheck the selected font, page size, margins, heading rules, table width, formulas, and page breaks in those downloaded files.
04Images, Mermaid, or formulas cause a failure
Use relative local image paths, valid Mermaid syntax, supported TeX delimiters, and shorter expressions when possible. Mermaid in ZIP packages needs the browser-rendered asset flow supported by the workspace.
05The download is missing or has expired
Download as soon as the job completes. If the file has already been cleaned up, start a new conversion from the reusable Markdown source instead of relying on an expired temporary link.
Next step