By leveraging the Page Processing feature in Parseur, you can enhance your document processing workflow by selecting specific pages for data extraction, saving some precious credits and improving AI data quality along the way.
Please Note: This feature is exclusively available for PDF documents and TIFF images. Text-based documents, including emails, are not applicable as they are treated as single-page documents.
Why use the page processing option?
Using the page processing option has the following benefits:
Save credits: often, only a portion of a document contains relevant data. The page processing option allows you to restrict parsing to only that subset of pages. Only the actual number of pages read will be deducted from your credit quota.
Improve AI quality: shorter documents often yield better AI extraction results. By only focusing on pages with relevant data, you optimize the context for AI processing, resulting in higher-quality extractions.
How do I access the page processing option?
Find this option under the Processing tab within your mailbox settings.
How does it work?
You can set your mailbox to process:
All pages: the default setting, processing every page.
Odd or even pages: choose to process either odd (1, 3, 5, etc.) or even (2, 4, 6, etc.) pages.
Specific pages or ranges: define individual pages (e.g.,
1
), ranges (e.g.,2-5
), or a combination (e.g.,1, 3-5
), separating selections with commas. Use brackets to refer to pages from the document's end (e.g.,(1)
for the last page,(2)
for the second-to-last page, etc.).
Examples:
1
: extracts data from the first page only2-5
: parses pages 2 through 5(1)
: extracts data from the last page only(3)-(1)
: extracts data from the last three page1, 3-5, (1)
: extracts from the first page, pages 3 to 5, and the last page
Impact on Credit Usage
Your credit consumption aligns with the actual number of pages processed. For instance, specifying the first page of a 100-page document only deducts one credit.
Template Considerations
Should a mandatory field reside on a non-processed page, you'll encounter a "New Template Needed" error.
Change the field to optional in the template or modify your page processing range to include this field to resolve this.
Impacts on metadata fields
The content of metadata fields (e.g., Content
, TextDocument
, HtmlDocument
) will be confined to the extracted pages as per your specified range.
โ