Uploading Documents

Add documents to BodhAI for indexing and retrieval.

Uploading Documents

BodhAI indexes your documents so they can be retrieved and cited in answers. Indexing is asynchronous — you can keep using the app while it runs.

How to upload

  1. Click Upload in the sidebar
  2. Drag files onto the upload area, or click to browse
  3. Choose where the documents go — one or more groups, or Personal (visible only to you and org admins). Only groups where you hold edit permission are offered as targets. See Groups & Access Control.
  4. Click Upload — the file transfers immediately
  5. A notification appears in the bottom-right corner: Indexing…
  6. When indexing completes, the notification changes to Ready — N chunks indexed with a link to the document

You can navigate away as soon as the upload confirmation appears. Indexing continues in the background.

Batch upload

You can select and upload multiple files at once. Each file is indexed independently. The notification system shows progress per file.

File size limit

Maximum 500 MB per file. For very large files (e.g. 200+ page PDFs), indexing may take 2–5 minutes.

What BodhAI does during indexing

  1. Parses the document into structured elements (text, tables, headings)
  2. Runs OCR on scanned pages and image files — printed text only (see Supported File Types)
  3. Splits content into semantic chunks at natural boundaries
  4. Generates embeddings for each chunk
  5. Stores chunks and embeddings in the local vector database
  6. Extracts named entities (people, locations, organisations, part numbers) and relationships

No document content ever leaves your network during this process.

Duplicate detection

If you upload a file with the same name and content as one already indexed, BodhAI returns a link to the existing document instead of re-indexing. Uploading an updated version of a file (same name, different content) creates a new document entry.

Supported file types

See Supported File Types for the full list.

If an upload fails

  • File too large — split the document or compress images before uploading
  • Unsupported format — convert to PDF before uploading
  • Indexing fails — check that the BodhAI server is running and the Python sidecar is healthy. See Troubleshooting.