Knowledge base
Upload documents to a dataset
Documents are the unit of content inside a dataset. Each file's text is extracted, split into chunks, embedded, and added to the index. Agents retrieve chunks at runtime.
What you'll learn
- Which file formats Dezifi accepts
- How to upload files
- What each indexing status means
- How to reindex, update and delete documents
Supported formats
Dezifi accepts PDF, TXT, MD, DOC, DOCX, HTML and CSV files, up to 10 MB each and 10 files per upload. JSON is not accepted. Text is extracted from each file and split into chunks of up to about 1,000 characters (200 overlapping, set by the search engine), whatever the format — a CSV row is not kept as a unit. Split larger files before upload.
Upload a file
- 1
Open the dataset
From Knowledge Base, click the dataset you want to load into. You'll see the document list — empty on a new dataset. Choose its indexing key first if you have not. - 2
Click Upload documents
Click Upload documents, then click to pick files (up to 10; multi-select and mixed formats are fine) and start the upload. Files are picked, not dropped: the dialog has no drop zone. - 3
Watch indexing progress
Each row shows its status: Queued, Processing, Indexed or Failed. The list does not refresh on its own — click Refresh. A small file indexes in seconds. The dataset is searchable as soon as the first document is Indexed. An Indexed document with 0 chunks yielded no text (for example a scanned PDF) and answers nothing. - 4
Handle failures
A Failed row shows the reason. Click Reindex on the row to try again (for example after fixing the indexing key), or delete it and upload a clean copy.
View and manage documents
- 1
Find a document
Search by file name, filter by status and sort by name, size, status or date. The list shows each document's type, size, status and chunk count; the chunks themselves are not shown in the dashboard. - 2
Update a document
Uploading a file with the same name adds a second document; it does not replace the first. Delete the old version, then upload the new one, or both stay searchable. - 3
Delete a document
Click the delete button on the row and confirm. Its chunks are removed from the index so agents stop retrieving them.
Frequently asked questions
- Can I upload a folder or zip?
- No. Zip files are not accepted and there is no folder upload. Pick the files themselves, up to 10 per upload; to load many files, repeat the upload or use dak rag ingest in a loop.
- How are PDFs with images or tables handled?
- Only the text is extracted; tables are read as plain text and split like everything else. Images are skipped and there is no OCR — a scanned PDF has no text to index. Export image-heavy content to a text-first format or run OCR first.
- Are my documents used to train any model?
- No. Uploaded documents are stored in your workspace's vector store and used only for retrieval by your agents. They are not used to train base models or shared across tenants.
- Why is one file still Queued?
- The list does not refresh on its own; click Refresh. If a file stays Queued for more than ten minutes, click Reindex on its row, or delete it and upload it again.