Uploads, waiting and downloads: troubleshooting
Find the next useful action without restarting a task unnecessarily.
A file is rejected before processing
Confirm the actual format is PDF, DOCX, XLSX, TXT or Markdown and the file is within 20 MiB. A PDF can contain up to 100 pages. Open the source locally to check it is readable and unlock password-protected content. If it is an old Office format, export a supported copy rather than merely changing the extension.
A task is queued or taking longer
The service shares processing capacity. Queue time is separate from the time needed to read pages and generate output. Scanned files and complex pages can take longer than short text. Keep the existing task view and check its status. Submitting the same file again creates additional work and does not make the original task finish sooner.
The browser loses its connection
A network interruption does not cancel a submitted task. Use the existing task’s status action or return in the same browser session. If you integrate through the API, resume polling the saved job ID. Uploads interrupted before completion may need another upload, but a task already accepted by the server should be checked first.
A result or key is unavailable
Use the browser session or key that owns the task. Another session cannot retrieve it simply by knowing its ID. Trial credentials expire, generating a new key revokes the old one, and terminal task data are cleaned after the retention window. Download completed results promptly. For integration troubleshooting, record the error code and time without including file content or secret keys.
Before you move on
- Check format, size and encryption first.
- Check an accepted task before resubmitting.
- Use the original owner’s credentials.
- Save completed results promptly.