Skip to content
Nappai Documentation

Extract Text from Binary

Extract Text from Binary reads files and extracts their text. You give it a document as it arrives (an email attachment, an uploaded file, a file downloaded from a cloud folder) and it returns its content as text, ready for an agent to read, summarise or search.

It works with the most common office formats: PDF, Word (DOC and DOCX), PowerPoint (PPTX), Excel (XLS, XLSX and XLSB), OpenDocument sheets and documents (ODS and ODT), CSV, Markdown, plain text and ZIP files. In images, it recognises the text written in them (in Spanish and English).

When to use it

  • Read PDF invoices that arrive by email, to extract amounts and dates afterwards.
  • Turn into text the contracts or quotes your customers upload to a shared folder.
  • Convert an Excel sheet into text so an agent can answer questions about it.
  • Read the text in photos of receipts or scanned documents.

Configuration

FieldWhat it does
Binary (required)The file or files you want to extract text from. Connect the file coming from another component.

This component supports Batch Mode, to process a list of files, and it can also be used as an agent tool. See Key concepts.

Output

The Data output returns the extracted text. Depending on the document type, one file can produce one or several results (for example, parts or pages of a long document).

If a file has no text to extract or its format is not recognised, it produces no result and the other files are processed anyway.

You can connect the result to an agent, to Entities extraction, to Summarizer or to any component that works with text.

Example

The purchasing department receives supplier quotes by email.

  1. Connect the email attachments to Binary.
  2. Send the extracted text to an agent with the instruction “List the supplier, the total amount and the validity date”.
  3. Store the answer in a spreadsheet to compare offers.

Good practices

  • With images, use sharp, well-lit photos: text recognition depends on quality.
  • If you only need to recognise text in images, Easy OCR gives you more control over languages.