Documents & recognition
The assistant can read the content of uploaded files, including scanned documents and images.
When a user uploads a file to iVendNext, the assistant can extract its content for further work, such as pulling figures off a scanned invoice or reading a supplier's price list.
What it can read
| File kind | What the assistant does |
|---|---|
| PDF documents | Extracts the text, and can pull out tables. |
| Images (scans, photos) | Reads the text using built-in document recognition. |
| Spreadsheets | Parses the rows and columns into structured data. |
| Word-style documents and text files | Extracts the text. |
Operations
| Operation | Use |
|---|---|
| Extract | Get the text or data content of any supported file. |
| Recognise (read images) | Read text from a scanned image or photo. |
| Parse data | Turn a spreadsheet into structured rows. Spreadsheets only. |
| Extract tables | Pull tables out of a PDF. PDFs only. |
Document recognition
Recognition works out of the box, supports more than 80 languages, and runs safely in an isolated process so a difficult document never affects the rest of the system. Administrators choose the recognition engine and default language on the Document Recognition tab (section 8), and a user can specify a different language for a one-off request.
Limits
| Limit | Value |
|---|---|
| Maximum file size | 50 MB |
| Maximum PDF pages per request | 50 (adjustable per request) |
| Recognition timeout | 120 seconds |
Last updated 9 hours ago
Was this helpful?