Add Documents
Upload files and follow them through the chunk-and-embed processing pipeline.
Open a knowledge base and use the Documents tab to upload files. You can drag and drop or browse, and you can select multiple files at once. Each file is uploaded, then processed in the background.
Supported file types
Text is extracted from common document and data formats:
- Documents — PDF, DOC, DOCX, TXT, MD, HTML
- Presentations — PPT, PPTX
- Spreadsheets / data — CSV, XLS, XLSX
Scanned PDFs and image-only pages are run through OCR before text extraction, so a PDF that is really a photo of a page still becomes searchable.
Storage limits
Coming soon. Storage limits are on the upcoming release track and are not yet enforced in the latest released version of ByteChef.
The total size of all knowledge base documents in a tenant is capped — by default at 1 GB.
Uploads that would exceed the cap are rejected with a limit-exceeded error until documents are
deleted. Self-hosted administrators can change or remove the cap with the
BYTECHEF_AI_KNOWLEDGE_BASE_MAX_SIZE_BYTES
environment variable (0 = unlimited).
The processing lifecycle
Every document moves through a small set of states, shown as a status on each row in the Documents tab:
| Status | Meaning |
|---|---|
| Uploaded | The file has been received and is queued for processing. |
| Processing | Text is being extracted, chunked, and embedded. |
| Ready | The document is fully indexed and searchable. |
| Error | Processing failed — hover the row to see the reason. |
A document only contributes to search results once it reaches Ready. Large files and OCR-heavy PDFs take longer; the status updates as processing completes.
Re-processing
Deleting and re-uploading a document re-runs the full pipeline with the knowledge base's current chunk settings. This is how you re-chunk older documents after changing those settings.
Tagging documents
Documents can be tagged individually. Tags are the unit of retrieval filtering: a workflow
or agent can be pointed at only the documents carrying a given tag. Tag consistently if you
plan to keep distinct content (for example policy, product, legal) in one knowledge
base but retrieve from each independently.
Keeping documents in sync automatically
Uploading files by hand is fine for static reference material. For content that lives in another system and changes over time — a database table, a CMS, a help center — configure a source instead, and ByteChef will keep the knowledge base up to date on a schedule.
How is this guide?
Last updated on