> ## Documentation Index
> Fetch the complete documentation index at: https://filla.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# PDF Text Extractor

> Extract text content from PDF attachments in Airtable, with OCR support for scanned documents.

## What this tool does

The PDF Text Extractor reads PDF files from an Airtable attachment field and writes their text content to a text field. It handles both digital PDFs (with embedded text) and scanned PDFs (image-based, requiring OCR).

## Settings reference

### Source

| Setting                  | Description                                                                 |
| ------------------------ | --------------------------------------------------------------------------- |
| **Source table**         | The table containing PDF attachments.                                       |
| **PDF attachment field** | The attachment field containing PDF files.                                  |
| **Output text field**    | A Single Line Text or Multiline Text field where extracted text is written. |
| **Enable OCR**           | Use Tesseract.js OCR for scanned or image-based PDFs.                       |

### Output settings

| Setting                      | Description                                                                            |
| ---------------------------- | -------------------------------------------------------------------------------------- |
| **Include page markers**     | Add `--- Page N ---` separators between pages.                                         |
| **Character limit handling** | Truncate at 100K characters (Airtable's limit) or Split across multiple fields.        |
| **Overflow fields**          | Additional text fields for the split strategy. Only shown when split mode is selected. |
| **Multiple PDFs handling**   | When a record has more than one PDF: Combine text from all, or Extract first only.     |

### Extraction settings

| Setting                         | Description                                                                                                               |
| ------------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
| **Extract specific page range** | Toggle to limit extraction to certain pages.                                                                              |
| **Start page**                  | First page to extract (1-indexed). Only shown when page range is on.                                                      |
| **End page**                    | Last page to extract. Only shown when page range is on.                                                                   |
| **OCR language**                | Language for OCR: English, Spanish, French, German, Chinese (Simplified), Japanese, or Arabic. Only shown when OCR is on. |
| **Generate keyword index**      | Extract the top 20 keywords by frequency from the text.                                                                   |
| **Keyword field**               | A text field to store the extracted keywords. Only shown when keyword extraction is on.                                   |

### Record filters

| Setting                            | Description                                              |
| ---------------------------------- | -------------------------------------------------------- |
| **Source view**                    | Limit to records in a specific view.                     |
| **Skip already processed records** | Skip records where the output field already has content. |

## How the tool detects scanned PDFs

The tool automatically checks each page for text content. If a page contains fewer than 50 characters, it's treated as a scanned/image page and OCR is applied (if enabled). Digital pages with text are extracted directly without OCR.

## Handling the 100K character limit

Airtable text fields have a 100,000 character limit. If a PDF's extracted text exceeds this:

* **Truncate at 100K**: Text is cut at the limit. The remaining content is lost.
* **Split across multiple fields**: Text is split at line boundaries across multiple fields you specify. This preserves all content.

## Common questions

<AccordionGroup>
  <Accordion title="Can it extract text from password-protected PDFs?">
    No. Password-protected PDFs are skipped with a warning in the execution log.
  </Accordion>

  <Accordion title="What if a PDF is corrupted?">
    Corrupted or invalid PDF files are skipped with an error in the execution log.
  </Accordion>

  <Accordion title="How accurate is the OCR?">
    Accuracy depends on the scan quality, resolution, and language. High-quality scans with clear text produce good results. Low-quality or handwritten scans may have reduced accuracy.
  </Accordion>

  <Accordion title="Can I extract from PDFs and then search the text in Airtable?">
    Yes. Once the text is in a text field, you can search and filter by it in Airtable's grid view.
  </Accordion>
</AccordionGroup>