Multi-page TIFF Files
Multi-page TIFFs are common in archival scanning (bound volumes, registers, photograph albums). Archivers.ai handles them in one of two ways — your choice per file.
The two strategies
1. Treat the TIFF as one item
The whole TIFF is one item in the accession. OCR runs across all pages and is concatenated. Vision analysis runs on the first page (the cover or title page) but reads context from the whole file.
Use this when the TIFF is intellectually a single thing — a bound volume, a sequential register, an album. The catalogue record describes the volume; pages aren't independently catalogued.
2. Split into one item per page
Each page becomes a separate image item in the accession. OCR, vision, and classification run independently on each page.
Use this when each page is intellectually distinct — a photograph album where each photo is its own item, a folder of loose papers scanned together.
How to pick
You're prompted during Step 3, Review Classifications, where a multi-page TIFF gets its own blue card — Multi-page TIFF (N pages) — offering Process as document (OCR all pages) or Split into N individual images. You can change your answer there right up until processing runs.
There's no right answer — it's a description decision. If you're cataloguing a bound register, treat it as one item. If you've scanned a photograph album, split it.
Implications for cost
Either way you pay by the page: a document costs 1 credit per page, and a split TIFF becomes N image items at 1 credit each. Splitting also counts N against the Community plan's 10-items-per-accession cap, where keeping it as one document counts 1.
What about PDF?
PDFs are handled similarly — the platform extracts text from all pages but treats the PDF as a single item by default. There's no automatic page-splitting for PDFs; if you need separate items per page, split the PDF before upload using your preferred tool.
What if I change my mind later?
Once the file has been processed the choice is made. More → Read the records again on the accession puts the files back through the pipeline, which spends credits a second time; it does not change a split back into a single document.