Support Center

Retrieval of Failed to Process Documents

Last Updated: Jan 08, 2016 03:22PM EST
NOTE: Redirecting to updated article at

Once a document is submitted, is there a guarantee that it will be processed? If not, can this be detected via the API?
A document that is returned with a “FAILED” status was not processed by the Semantria engine. FAILED documents will be delivered without any analytic information. Semantria guarantees that every document will be delivered back to the client regardless of whether or not it was successfully processed or FAILED. A document can fail to be processed if it doesn't meet minimum Semantria requirements, including:

  • Alphanumeric characters threshold has been exceeded (the default is 80%)*
  • The Document's sentences are too long. By default the sentence length is 1000 words per sentence.**
  • White spaces aren’t available, so the text is just one word.
  • The analysis execution time exceeds 120 milliseconds per method (entities, queries, categories, etc.)***
You can check the status of any document within 24 hours of being queued using the "Get document" API end-point if needed. All documents that have not yet been retrieved will remain on server for 24 hours. If they are retrieved, the data will be deleted from the servers within 15 minutes.

*The alphanumeric characters threshold is adjustable through the configuration -
**The allowed sentence length can be extended on demand by contacting However, increasing the allowed sentence length will only trigger the engine to accept the document for processing. The engine will not take into account the +1000 word sentence in its evaluation. Therefore, best practices recommend to use punctuation appropriately throughout a document.
***Method execution time can be adjusted by contacting
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found