For ManualTrigger Automate, effortlessly convert bank statements into markdown format using advanced vision language models. This workflow downloads PDFs from Google Drive, splits them into images, and transcribes the content while preserving tables and details. Extract key deposit data efficiently, ensuring sensitive information is handled securely. Ideal for automating document processing and enhancing data extraction accuracy.
- Financial Analysts: Professionals who need to analyze bank statements quickly and accurately.
- Accountants: Individuals looking to streamline their accounting processes by automating data extraction from bank statements.
- Small Business Owners: Entrepreneurs who manage their finances and require clear insights from their bank statements.
- Data Scientists: Analysts interested in working with document processing and automation workflows that utilize language models.
- Developers: Tech-savvy individuals who want to integrate document processing capabilities into their applications using n8n and LangChain.
This workflow addresses the challenges of extracting and analyzing data from bank statements, particularly those in PDF format. It automates the process of converting scanned PDFs into structured markdown text, enabling users to efficiently identify and extract key data such as deposits and withdrawals. This reduces manual effort, minimizes errors, and enhances the speed of data processing, making financial analysis more efficient.
1. Manual Trigger: The workflow starts when the user clicks the ‘Test workflow’ button.
2. Get Bank Statement: Downloads the bank statement PDF from Google Drive using its file ID.
3. Split PDF into Images: Sends the PDF to a service that converts each page into separate images.
4. Extract Zip File: Unzips the zip file containing the images of the PDF pages.
5. Images To List: Converts the unzipped images into a list format for further processing.
6. Resize Images For AI: Resizes the images to 75x75 pixels to optimize them for processing by the AI model.
7. Transcribe to Markdown: Uses a language model to transcribe the images into markdown format, capturing all text and table structures.
8. Combine All Pages: Combines the transcribed text from all pages into a single document.
9. Extract All Deposit Table Rows: Extracts specific deposit data from the combined markdown text, structuring it for easy analysis.
10. Output: The final output includes structured data that can be used for financial analysis.
- File ID: Change the file ID in the ‘Get Bank Statement’ node to point to a different bank statement stored in Google Drive.
- Image Processing Settings: Adjust the image resizing parameters in the ‘Resize Images For AI’ node to optimize for different AI models or specific use cases.
- Model Selection: Swap out the Google Gemini model for another language model that supports image inputs if preferred, ensuring compatibility with the workflow.
- Data Extraction Logic: Modify the prompts and extraction logic in the ‘Extract All Deposit Table Rows’ node to fit different formats or types of bank statements.
- Add Additional Steps: Users can expand the workflow by adding more nodes for additional data processing or integration with other applications.