How to Summarize Scanned PDFs with ChatGPT Using PDF.co and Zapier

6 Minutes Read

This tutorial creates a Zap that converts scanned PDFs into searchable documents, extracts their text, and generates summaries with ChatGPT.

Prerequisites

You will need:

  • A Zapier account
  • A PDF.co account
  • A ChatGPT (OpenAI) connection in Zapier
  • A Google Drive folder for scanned PDFs

Step 1: Create the Google Drive Trigger

Create a Zap and select:

  • App: Google Drive
  • Trigger event: New File in Folder

Connect your Google Drive account and select the folder containing incoming scanned PDFs.

Test the trigger with a sample scanned PDF. Map the Google Drive File value in subsequent PDF.co steps.

Step 2: Make the PDF Searchable

Add another step:

  • App: PDF.co
  • Action: Convert Scanned PDF to Searchable

Connect your PDF.co account by signing in and authorizing Zapier.

Configure the action:

  • Source PDF URL: Map the PDF file from Google Drive.
  • OCR Language: Select the document’s language, such as English.
  • Pages: Leave empty to process every page.
  • Output PDF Name: Enter a name such as searchable-document.pdf.

Test the step. PDF.co will apply OCR and return a searchable PDF URL.

If the searchable PDF is also needed as a deliverable, add a storage step later in the Zap and save this URL to Google Drive or another permanent destination.

Step 3: Extract the PDF Text

Add another PDF.co step:

  • App: PDF.co
  • Action: PDF to Text

Configure it as follows:

  • Source PDF URL: Map the searchable PDF URL from the previous step.
  • Pages: Leave empty to extract all pages.
  • OCR Language: Select the document’s language.
  • Inline Output: Select True, if this field is available.

Test the action and confirm that the response contains the extracted text, commonly returned in the Body field.

Additional extraction settings are described in the PDF.co PDF to Text documentation.

Step 4: Summarize the Text

Add a new action:

  • App: ChatGPT (OpenAI)
  • Action: Analyze Text

Configure the action:

  • Text to Be Analyzed: Map the extracted PDF.co text.
  • Model: Select an available model.
  • Analysis Type: Select Summarize.

If the action provides additional instructions, enter:

Summarize this document clearly. Include its main subject, key points, important dates or figures, and any actions or decisions mentioned.

The current Analyze Text action is designed for summarization and avoids using the legacy Conversation action. Available ChatGPT actions are listed on the ChatGPT Zapier integration page.

Step 5: Use the Summary

Test the ChatGPT step and review the generated summary.

Add another action to send or store the result. For example, you can:

  • Save it in Google Docs
  • Add it to a Google Sheets row
  • Send it through Gmail or Slack
  • Store it with the searchable PDF
  • Add it to a database or document-management system

Step 6: Publish the Zap

Test the complete workflow with another scanned PDF. Confirm that the file becomes searchable, its text is extracted, and ChatGPT generates an accurate summary.

Publish the Zap when the workflow succeeds.

Conclusion

The Zap now monitors a Google Drive folder, converts scanned PDFs into searchable documents using PDF.co, extracts their text, and sends that text to ChatGPT for summarization. The searchable PDF and generated summary can both be retained for later use.

Related Tutorials

See Related Tutorials