Document processing status
After you upload a document, Zahen processes it automatically before it becomes searchable. You can track progress in the document list under Administration → Document Upload.
Processing stages
Section titled “Processing stages”| Status | What’s happening |
|---|---|
| Pending | The file has been received and is queued for processing. |
| Parsing | Zahen is reading the file and extracting its text content. |
| Embedding | The extracted text is being split into chunks and converted into searchable vectors. |
| Ready | Processing is complete. The document is now searchable and can be cited in answers. |
| Failed | Something went wrong. The document is not available to the assistant. |
You don’t need to wait for processing to finish — the stages typically complete within a minute or two for most documents, depending on file size.
When a document shows “failed”
Section titled “When a document shows “failed””A failed document is not searchable and will not appear in any answers. The most common causes are:
- Image-only PDF — a scanned document where the pages are images with no extractable text. Zahen needs selectable text to work with.
- Corrupt or incomplete file — the file was damaged before or during upload.
- Unsupported format — a file extension mismatch or a format not in the supported list.
To resolve a failure:
- Open the original file on your computer and check that you can select and copy text from it.
- If it’s a scanned PDF, run it through an OCR tool to produce a text-layer version, then save as PDF or export as DOCX.
- If the file appears intact and in a supported format, try re-saving it from its source application.
- Re-upload the corrected file.
Removing or replacing a document
Section titled “Removing or replacing a document”If a policy is outdated or incorrect, remove it (or replace it with an updated version) promptly. The assistant will continue citing any document that has “ready” status until it is removed — there is no automatic expiry. See Writing documents that answer well for more on keeping the knowledge base current.