The Document Splitter feature lets you break large documents into smaller, structured sub-documents.For example, if you upload a 40-page file that contains multiple document types (invoices, receipts, IRS documents, custom models, etc.), you can split it into smaller sections and send each type into the correct flow automatically. This saves time and ensures that each document type is processed with the right model.
Here's how it works:
Go to your Custom Model.
Choose Document Splitter under Model Category in the Info tab.
Go to Splitter tab.
Define your splitting rules:
Name: A unique name to identify the rule (used for classification on sub-documents). Rule: Choose startsWith or endsWith (explained in the Value section). Value: You can use either plain text or a RegEx (regular expression) as the rule value to identify sub-documents. The splitter will check for matches based on the selected rule type (startsWith or endsWith).
Example: If you set the value to page 1 of and select startsWith, every time the splitter finds page 1 of anywhere on a page, it will create a sub-document starting with that page. It’s not necessary for the text to be at the beginning of the page.
If there’s no matching endsWith rule, the splitter will continue grouping pages together until the next match of another rule.
Model type: Select which model (or default classification) should handle this sub-document.
Target flow: Decide which flow this sub-document should be uploaded into.
Assign the custom model to an ingestion flow (this activates the splitting — only OCR is applied, no extra extraction).
Upload your document to the ingestion flow and watch it split into the defined sub-documents.