Who this is for
Someone responsible for an office archive of printed business documents that has already been scanned to PDF but cannot be searched, and who needs to find documents by their words.
People open scan after scan looking for one name or phrase, because searching the folder finds nothing inside the files.
The result
You receive a searchable copy of each PDF in the folder, with the original untouched, and a report on the search terms you chose: which pages they were found on, which expected pages missed them, with the page image shown for every miss, and which pages had poor recognition. The report is held to a pass floor agreed in writing before processing, so you know what a search can and cannot find.
What is included
- Up to 500 pages of already-scanned PDFs in one or two named languages, with no personal or confidential content; we do not scan paper
- Assess the scans for resolution, skew and blank pages before processing, class each page as clean (printed text, upright within an agreed angle, at least 300 dpi, not faded) or other, and tell you what is likely to recognise poorly
- Agree a pass floor in writing after that assessment and before processing
- Add a text layer to copies, keeping the originals, in an archival PDF format if you ask for it
- Run a search test with 15 to 20 terms you know appear on known pages, and list pages where recognition was poor
You receive
- A searchable copy of each PDF in the batch, with the original kept
- A scan assessment that classes each page as clean or other, and the pass floor agreed from it
- A search-test report: each term, the page you expected, whether it was found, and the page image for every term that was not
- A list of pages with no or poor recognised text, with a reason where we can give one
What is not included
- Scanning paper, or rescanning poor pages
- Recognising handwriting
- A promised recognition accuracy for the batch as a whole, or retyping and correcting the recognised text
- Redaction, legal admissibility, or advice on records and retention
- Extracting data into a spreadsheet
- Documents holding personal, confidential or regulated information
What we need from you first
- The number of pages, the language, the era or type of the documents and the resolution of the scans if known
- Whether you can select any text in the PDFs now
- Five or six words you would search for; no documents in the first enquiry
Never send passwords, keys, customer records or confidential code in the first enquiry. Secure handover is agreed after scoping.
How we check it is done
- Every input PDF has a searchable output copy, or is listed with the reason it has none, and the page count of each output equals that of its original.
- For each of your 15 to 20 agreed search terms, the report states whether the term was found on the page where you know it appears and shows the page image for every term not found. The result meets the pass floor agreed in writing before processing: every term whose known page the assessment classed as clean is found on that page, and at least the number of terms agreed in writing for the other pages is found.
- Pages whose recognised text is empty or below the agreed threshold are listed with their file and page number.
- The original files are identical before and after the job.
You run some of your own searches in the copies, read the report and check the page images for any misses, then sign off in writing. Payment follows sign-off.
When we would stop or decline
- The pages are mainly handwriting, or the scans are too poor to read
- The documents contain personal, confidential or regulated information
- You want a guaranteed recognition accuracy
- The scan assessment shows that no honest pass floor can be set; we say so and nothing is charged
Questions
How accurate will the recognised text be?
We do not promise an accuracy. We test the terms you choose, show every miss with its page image, and list pages that recognised poorly. Before processing we agree in writing a pass floor, and you do not pay if the result falls short of it.
What if the scans are too poor for any floor to be set?
We tell you after the scan assessment, before any processing, and nothing is charged.
Can you read handwriting?
No. Handwritten pages are outside this job and will be listed.
Do you scan the paper?
No. We work on PDFs you have already scanned.
Will it be in PDF/A?
If you ask for it, yes. A level B PDF/A file made from page images does not necessarily support indexing of its text, which is why we run a search test whichever format you choose.
Price and terms
From £195 · untested offer price. The fixed price is confirmed in writing and paid after the agreed checks pass, including the pass floor, and you sign off. No payment before sign-off.
This is a new service with no published client results. The price is a starting point we have not yet tested with buyers. Nothing is ordered or charged by the enquiry. The full specification is on the Synthetic Industry catalogue.