Process Large Documents with OCR using SubworkflowAI and Gemini

This n8n workflow demonstrates a robust solution for processing large documents, extracting text via OCR, and then summarizing or analyzing the content using Google Gemini. It's designed to handle documents stored in Google Drive, making it ideal for automating data extraction and content analysis from scanned PDFs, images, or other non-text-searchable files.

What it does

This workflow automates the following steps:

1. Manual Trigger: Initiates the workflow upon a manual click, allowing for on-demand processing.
2. Google Drive File Listing: Retrieves a list of files from a specified Google Drive folder.
3. Filter for Documents: Uses an "If" node to filter the retrieved files, likely based on file type (e.g., PDF, image) or a specific naming convention, to ensure only relevant documents are processed.
4. Split Out Files: Processes each filtered document individually, ensuring that subsequent steps operate on one document at a time.
5. SubworkflowAI OCR: Sends each document to a SubworkflowAI API endpoint for Optical Character Recognition (OCR). This step extracts all text content from the document, including from scanned images.
6. Wait for OCR Completion: Pauses the workflow execution, waiting for the OCR process from SubworkflowAI to complete. This is crucial for handling asynchronous OCR operations.
7. Google Gemini Analysis: Once the OCR text is available, it's fed into the Google Gemini node. This node can then perform various AI tasks, such as summarizing the document, extracting key information, answering questions about the content, or generating further insights.

Prerequisites/Requirements

To use this workflow, you will need:

• n8n Instance: A running n8n instance (cloud or self-hosted).
• Google Drive Account: Configured with n8n credentials to access your documents.
• SubworkflowAI Account/API Key: An account with SubworkflowAI (or a similar OCR service) and the necessary API key/credentials to perform OCR operations. The workflow uses an HTTP Request node, implying a direct API call to SubworkflowAI.
• Google Gemini API Key: Access to the Google Gemini API, configured as a credential in n8n.

Setup/Usage

1. Import the Workflow:
   • Download the provided JSON file for this workflow.
   • In your n8n instance, go to "Workflows" and click "New".
   • Click the "Import from JSON" button and paste the workflow JSON or upload the file.
2. Configure Credentials:
   • Locate the "Google Drive" node and configure your Google Drive OAuth2 credentials.
   • Locate the "HTTP Request" node (for SubworkflowAI) and set up the necessary authentication (e.g., API Key in headers or query parameters) according to SubworkflowAI's documentation.
   • Locate the "Google Gemini" node and configure your Google Gemini API Key credential.
3. Customize Nodes:
   • Google Drive: Adjust the "Folder ID" to point to the specific Google Drive folder containing the documents you want to process. You might also want to refine the file search parameters.
   • If: Review and adjust the conditions in the "If" node if you need to filter documents based on different criteria (e.g., file name patterns, specific MIME types).
   • HTTP Request (SubworkflowAI):
     • Update the URL to your specific SubworkflowAI OCR endpoint.
     • Ensure the Body Parameters correctly send the document data (e.g., file URL or binary data) required by SubworkflowAI.
     • Verify the Headers for any API keys or authentication tokens.
   • Wait: The wait time might need adjustment based on the typical processing time of your SubworkflowAI OCR tasks.
   • Google Gemini: Customize the prompt and model parameters in the "Google Gemini" node to define how you want to analyze or summarize the OCR'd text.
4. Activate the Workflow: Once all configurations are complete, save and activate the workflow. You can then execute it manually using the "When clicking ‘Execute workflow’" trigger node.