We have two automations that can help you extract text from attachments:
Extract Text From PDFs: This is used exclusively for PDF files and can only extract actual text (i.e. not text from images within a PDF, for example when documents are scanned as pictures). It reads PDFs from one attachment field and combines the extracted text into one output per record, limited to the first 100,000 characters. To extract text from PDFs spread across multiple fields, you'd either need multiple instances of this automation, or you'd need to first combine the attachment fields into one.
Analyze and Modify Airtable Data with ChatGPT: This automation uses OpenAI's Responses API and is very powerful when it comes to analyzing attachments. It can handle multiple attachment fields at a time and supports image files (e.g. png, jpg, gif, and webp), PDFs, Word files (.doc, .docx), text files and several others.