Where the PDF can come from
- A link that starts with
https:// - An email attachment in Gmail, once your Gmail is connected
- A file in Google Drive, once your Drive is connected
How it comes back
Your AI gets the text with a heading for each page, so it can tell you exactly where something is. Long PDFs go to your workspace so your AI can search and read them in sections. You can also ask for specific pages:What it can’t read
- Scanned PDFs. Arora reads the text inside a PDF. A scan is a picture of a page with no text inside, so there’s nothing to read. Arora tells your AI when that’s the case instead of guessing.
- Text inside images in an otherwise normal PDF.
Limits
Technical details
Technical details
Tool name:
bridge_read_pdf. Sources: Gmail attachment, Google Drive file, or HTTPS URL. Returns markdown with a ## Page N heading per page. Results over 10,000 characters are staged to the workspace; a pages range (for example 3 or 2-8) always returns inline.