Pošlji račun OCR
Paper invoices prepared for a combined PDF file
Practical invoice OCR guide

Split a PDF With Multiple Invoices Without Losing Continuation Pages

·6 min read
← All guides

A PDF containing several invoices must be split into complete documents rather than isolated pages. Keeping continuation pages together preserves line items, totals and payment details for review and accounting delivery.

Split by invoice, not automatically by page

A page-by-page splitter works for single-page invoices. It can break a multi-page invoice by separating its line items from the final totals. Identify document boundaries from the invoice content.

For example, an eight-page PDF might contain four invoices covering page 1, pages 2–4, pages 5–6 and pages 7–8. The correct result is four files.

InvoicePDF pagesOutput
A1Single-page invoice
B2–4One complete three-page invoice
C5–6One complete two-page invoice
D7–8One complete two-page invoice

Review and confirm the split

Check the selected company and its delivery mode before uploading a batch. Open PDF batches to follow the job and review any uncertain boundaries.

  1. Upload the combined PDF for the correct company.
  2. The system processes the pages and proposes invoice boundaries.
  3. When a boundary is uncertain, compare the proposed ranges with the original.
  4. Correct each invoice’s first and last page as needed. Ranges must cover every page without gaps or overlap.
  5. Confirm the split, then check the resulting documents and their delivery status.

Recognise continuation pages

A repeated supplier logo does not necessarily start a new invoice. Look at invoice numbers, page numbering, continued tables and the totals section. Page headers and white space alone do not establish a reliable boundary.

Review specifications, delivery notes and attachments carefully. Keep each supporting page with the invoice it belongs to so the accounting reviewer receives a useful document.

Separate PDFs and structured extraction are different modes

The separate-PDF delivery mode splits and sends the original document sections without extracting accounting fields for structured export. OCR may still be used to identify invoice boundaries.

Choose structured processing when you also need suppliers, dates, VAT, totals and line items. Review the extracted result before using XML or API delivery. Select the mode according to what the receiving accounting system accepts.

Test delivery before processing a large batch

Start with a smaller sample when the document layout or destination is new. Confirm that each output contains the entire invoice and arrives under the correct company.

  • Each invoice contains every continuation page.
  • Page order, attachments and totals are preserved.
  • The company and destination email match the documents.
  • No unresolved boundary review or quota warning remains.
  • The archive confirms delivery rather than only upload completion.

Estimate OCR capacity by pages

A PDF with 30 invoices is not necessarily 30 pages long. Count the pages that will be submitted to OCR, including continuation pages, when estimating the allowance you need.

Paid plans share a monthly OCR page allowance. An uploaded file may still be waiting for review, processing, an allowance renewal or an upgrade. Check the job status before treating the batch as fully delivered.

Frequently asked questions

Should every PDF be split into individual pages?

No. Split it into complete invoices, keeping all pages of the same invoice together.

What if two invoices were combined?

Review the proposed page ranges in the PDF batch and correct the boundary before confirming the split.

Does separate-PDF delivery avoid all OCR?

Not necessarily. OCR can still identify the boundaries, even though line items and other accounting fields are not extracted for structured export.

Why was the PDF uploaded but not sent?

Upload is only the first stage. Check the batch stage, unresolved review, remaining allowance and any delivery errors.

Try the workflow on your own invoice

Free: 3 invoices and 10 OCR pages per month. Review the result before sending it to accounting.

FAQ

Common questions about invoice OCR

Quick answers about scanning, invoice data extraction, accounting integrations and multi-company workflows.

What is invoice OCR and which fields can it extract?+

Invoice OCR turns a PDF or image into structured accounting data such as invoice number, dates, supplier details, VAT ID, IBAN, net amount, VAT, gross total, VAT breakdown and line items.

Does invoice OCR work with Slovenian and English invoices?+

Yes. Slikaj Račun is designed for Slovenian and English invoices and common international invoice labels, so local and foreign supplier invoices can use the same workflow.

Do I need to replace my accounting software?+

No. Slikaj Račun complements the accounting system you already use. Keep your current invoicing and bookkeeping software while Slikaj Račun handles incoming-invoice capture, OCR, review and delivery.

Does it work with Minimax, Birokrat, Pantheon and other systems?+

Yes. You can forward the original invoice to an accounting import email. Approved structured data can also be used in UBL, eSLOG or API workflows depending on the destination system and plan.

What happens when OCR confidence is too low?+

The invoice is sent to review. The original document and extracted fields are shown side by side so a user can correct and approve the result before it continues downstream.

Can accounting firms manage multiple companies?+

Yes. Higher plans support multiple companies under one user account, separate delivery destinations and bulk document upload for faster processing.