Admin Access

Document Ingestion
Invalid admin key. Please try again.
The same admin key used for the Settings page.
Verifying access…
··

Ingestion

Document Ingestion
Settings

Document Library

Upload documents, paste text, or import transcripts from YouTube and podcasts. Everything you add is chunked, embedded, and stored in the vector database that powers this assistant.

Add Document

Drop files or a folder here

Select multiple files, or drag a whole folder. Each file is ingested as its own document.

PDF DOCX TXT MD EML
0 files selected
Transcript text will be chunked and embedded. The URL and media type are stored with every chunk so retrieval can link back to the source.
📚 Batch mode: title and source will be auto-derived from each filename. All files will be filed into the folder selected below.
Files are organized by folder so you can group sermons, books, articles, etc.
The document's own date — used by Year › Month grouping. Leave blank to extract from the title/filename when possible.
NVIDIA embeddings cap at 512 tokens. A 600-char hard ceiling is enforced — any chunk exceeding it is split further automatically.
Extracting text… 0%
—
Folders
—
Documents
—
Chunks
—
Collection

Ingested Documents

Loading documents…

Edit Document Metadata

Changing this renames the document's unique key. It must stay unique.
The document's own date. Used when grouping by Year › Month (or Testament › Book › Year › Month). If left blank, a date will be extracted from the title or filename when possible.

Document

Loading chunks…