PDF Part Number Splitter + Excel Audit

Turn mixed multi-page PDF reports into neatly split documents by part / service / reference number.

Upload up to 8 text-based PDFs and get a single ZIP containing separate PDFs and a beautifully formatted Excel audit workbook. Built for transport schedules, dispatch reports, work orders, service reports, logistics documents and back-office workflows.

How it works

1. Upload your PDFs directly, or provide HTTPS links to downloadable PDF files.
2. Pick how to handle unlabeled pages: attach to previous part (marked as inferred) or send to review.
3. Run. Find RESULT_ZIP and CONTROL_PARTES_XLSX from the run's Output tab (or under Storage > Key-value store).
4. Check Review alerts in the Excel workbook before distributing outputs.

Each identified reference becomes one PDF. Pages belonging to the same reference are reunited even when the source report is interleaved.

Input examples

Simple upload: use Upload PDF files in the input form. No coding needed.

Or submit JSON:

{
  "pdfUrls": ["https://example.org/path/to/service-report.pdf"],
  "unknownPagePolicy": "inherit"
}

You can customize part-number matching. For example, when each page contains ORDER ID: ABC-123456:

{
  "pdfUrls": ["https://example.org/path/to/orders.pdf"],
  "referenceRegex": "(?im)^ORDER\\s+ID\\s*:\\s*([A-Z0-9-]{3,35})"
}

Outputs

• RESULT_ZIP: PDFs grouped by reference + Excel control workbook.
• CONTROL_PARTES_XLSX: Excel workbook with sheets: Summary, Parts, Pages, Instructions (labels in Spanish).
• SUMMARY: machine-readable JSON with counts, errors, and review flags.
• Dataset: one row per successfully processed input PDF with at least one identified part/reference.

Supported formats and limitations

• Up to 8 PDFs per run, 25 MB per PDF, 200 pages per PDF, 500 pages overall.
• Digital, text-selectable PDFs only. Scanned PDFs without OCR layer are not supported.
• Pages with multiple different references are placed into a REVIEW PDF; never silently assigned.
• Unlabeled continuation pages may be inferred; this is not guaranteed, and must be manually verified.
• Two separate documents printed on the same source page cannot be split into independent pages automatically.
• User-supplied custom regex supports a first captured group or a group called ref.

Privacy

Only upload files you have a right to process. Original PDFs and generated outputs are stored in your Apify-run-managed storage. Check privacy, retention, and access settings. PDF text is processed in the Actor process, without third-party AI APIs.

Billing note

The Actor is compatible with Apify's built-in apify-default-dataset-item pay-per-event pricing: one emitted dataset row per input PDF successfully processed with an identified part number. Input files with no detected parts can still produce review files, but no chargeable dataset row. Prices are determined by the publisher, not hard-coded here.

Version 0.1. Test with non-sensitive examples before using on critical or confidential workflows.