Upload documents
Supported formats, the 25 MB cap, scanned PDFs, and choosing a chunking method per batch.
Open a knowledge base. The drop zone sits above the documents table; click to browse or drop
several files at once. Needs document.create.
Formats and limits
| Formats | .pdf .docx .txt .md .html .htm .csv .tsv .json .eml |
| Size | 25 MB per file |
| Not readable | .xlsx (export CSV/TSV first), .pptx (export PDF first), images, audio |
A scanned PDF will fail. Ragenta reads the text layer only. A scan is an image with no text to read — there is no OCR in the indexing pipeline yet. Run OCR elsewhere and upload a text PDF.
"Chunk this batch as"
Before dropping files, the Chunk this batch as picker overrides the chunking method for this batch only. Knowledge base default uses the knowledge base's setting. Ragenta refuses at upload time when a method cannot read that format (for example Table accepts only CSV/TSV).
After the upload
Each file becomes a row in the table at pending, then moves through the indexing steps — see
Document status. The table refreshes itself while any file is still
processing; you can leave the page.
Indexing costs credits (the ingestion and embedding operations on the Usage page). Uploading the same file with the same content again: Ragenta recognises it and reuses the processed parts — see Document detail.