Divideix per text

When a PDF is a batch export containing multiple documents concatenated together — hundreds of invoices in a single file, a day's worth of scanned forms, or an auto-generated report where each section starts with a known heading — splitting it by a fixed keyword is far faster than manually identifying page ranges. Split by Text scans each page for a phrase you specify and creates a new output file every time that phrase appears, effectively using the document's own content as its split map. This is the tool that completes accounts payable automation, form processing pipelines, and bulk document distribution workflows. An accounting system exports 500 invoices as a single PDF; Split by Text finds 'INVOICE NUMBER' on each separator page and produces 500 individual invoice files. A medical records system batches patient letters; the tool splits at 'Dear Patient' to produce one letter per patient. A logistics company receives daily manifests where each shipment begins with a barcode label containing 'SHIPMENT ID'; the tool isolates each shipment into its own file for downstream processing. You can choose whether the keyword page itself is included in the output file or discarded — useful for separator pages that carry no meaningful content of their own. Case-insensitive matching ensures you do not need to worry about capitalisation variations in auto-generated documents. Regular expression patterns are supported for advanced use cases where the split marker is variable, such as 'Invoice #\d+' matching any invoice number. Split by Text complements Split by Bookmarks for documents that lack a formal outline but have consistent textual markers instead. If your documents have both, bookmarks are usually more reliable since they are structural rather than content-based. For maximum flexibility, combine the two approaches: split by bookmarks at the chapter level, then split by text within chapters to isolate individual records. All file processing occurs on dokk.ai's secure infrastructure. Files are deleted within 60 minutes and never used for machine learning or shared with third parties. The output files are standard PDFs compatible with every reader, printer, and document management system.