How can you efficiently anonymize thousands of files, including multipage PDFs and scans? We explain the challenges of mass document anonymization and how Mycroft Sweeper PRO helps address them.
A bulk anonymization task may involve thousands of files, each containing hundreds of pages. This scale can arise when preparing procurement records for disclosure, medical documentation for research, or large collections of HR documents for analysis outside the organization. In these situations, opening a PDF and automatically redacting every detected name or address is not enough. The entire collection needs to be processed within a reasonable time, with appropriate rules for different documents, progress tracking, and a convenient way to review the results.
This is the difference between anonymizing a single file and bulk document anonymization. At scale, the entire workflow matters, from adding files and analyzing their contents to handling exceptions and creating the output documents.
When looking for a suitable solution, you may encounter terms such as “bulk anonymization”, “mass document anonymization”, and “anonymizing multiple documents”. In this article, they refer to processing a larger collection of files within a single workflow. “Batch anonymization” emphasizes the technical approach of processing documents in batches.
What matters is what the software can actually do with that collection. It should read the contents, including scanned text, identify data, apply the configured rules, and show which files are ready and which need review. Processing PDFs and scans is also a different task from anonymizing records in a database.
Knowing that an application supports PDFs does not tell you whether it can handle a scanned document containing hundreds of pages, or what happens when there are thousands of such files. The ability to process large numbers of pages, the time required for OCR and analysis, and the organization of work that may take many hours all matter.
Mycroft Sweeper PRO is designed for working with large document collections. Files enter a project queue and are processed sequentially, without having to open each one manually. The application also analyzes scanned PDFs and images without a text layer. It retains analysis results, so opening a document for review does not require running OCR again.
The time needed to process a particular collection depends on document length, scan quality, and the computer’s capabilities. Before starting a project involving thousands of documents, it is therefore worth checking processing speed using a representative sample of your own materials.
The information to be removed depends on the document and the purpose for which it is being prepared. Removing every detected value from a particular category may also hide information that needs to remain visible.
For example:
An employment contract may need to be prepared with the employee’s personal data hidden while the employer’s details remain readable.
Application forms may require selected contact details or identification numbers to be redacted while retaining the information needed to assess the application.
Reports may require information identifying individual participants to be removed while preserving findings and conclusions.
The organization preparing the documents decides which information should be hidden. The software should allow those decisions to be applied consistently across the entire collection.
Mycroft Sweeper PRO organizes work into projects. Each project is a separate workspace where the user adds documents and defines the processing rules: which information the application should detect and remove, when a file should be sent for review, and which method should be used.
This may involve an overlay specifying fixed areas of a form, or templates that allow the application to recognize a document type and apply the rules prepared for it. In projects using templates, recognition takes place separately on each page, including when a single PDF contains scans of different documents.
Many organizations regularly process recurring forms, applications, declarations, and certificates. Certain fields are known in advance to require redaction, such as personal contact details or identification numbers. A field should be redacted even when the form has been completed by hand and its contents cannot be read accurately.
For forms with a fixed layout, Sweeper PRO can use a redaction overlay. It specifies areas of the page to be covered regardless of the text recognized within them. When the position of a form varies between scans, aligning the overlay correctly becomes particularly important.
A document template can be used when the application needs to recognize the document type and locate its fields despite changes in layout. It does not rely solely on fixed field positions. It can also work with handwritten forms or documents where longer answers entered on a computer have shifted subsequent fields. We prepare templates for specific documents as a separate service for customers.
When working with thousands of files, users need to be able to start processing, monitor progress, and quickly find documents requiring action. A simple “analysis in progress” message is not enough. They need to know which files are waiting, which are ready, which require review, and which encountered an error.
Sweeper PRO organizes work into projects. Users add documents to the queue and start processing. Status and progress information are available for individual files, while search and sorting help locate a particular document within a large collection.
Work can be spread across several sessions. The application saves the analysis and review state, allowing users to return to an unfinished task. If the selections in a particular document need to be changed, it can be opened in the editor to leave selected information visible or manually mark an additional area for redaction.
Bulk anonymization becomes difficult to manage if every file has to be opened manually just to determine whether a human decision is needed. Review rules should be adapted to the document types and the purpose for which they are being prepared.
Sweeper PRO can send all files for review or only those containing uncertain detections. It is also possible to process files without reviewing every document beforehand. Files requiring a decision are sent for review, and their statuses distinguish them from completed documents and files that could not be processed.
Users can examine the application’s findings, make corrections, and approve the results. For large collections, it is also worth checking a representative sample of completed documents, covering different file types and less typical cases.
The pricing model matters when processing thousands of files. Charging for every page or analysis would make the total cost harder to estimate, particularly when some documents need to be processed again.
Mycroft Sweeper PRO does not charge based on the number of pages, files, or other measures of usage. OCR, analysis, and anonymization take place locally on the user’s computer, without sending document contents to an external cloud service. This allows organizations to plan work on large collections without an additional charge for every processed page.
Before processing thousands of files, it is worth assembling a sample that reflects the actual task: short and multipage PDFs, scans, handwritten forms, and documents requiring different anonymization rules. This sample can be used to establish which data should be hidden, select project settings, and check processing speed and the review workflow.
Effective bulk anonymization requires software that can handle both data detection and the scale and organization of the work. Mycroft Sweeper PRO combines processing large collections with projects, progress monitoring, selective anonymization rules, and tools for reviewing documents.
Can multiple PDFs be anonymized without opening each one separately?
Yes. In Mycroft Sweeper PRO, files are added to a project and processed sequentially according to its settings. Documents sent for review can then be checked in the editor.
Does bulk anonymization support scans and multipage documents?
Yes. OCR enables the application to read scanned text, and a project can contain many files, including documents with hundreds of pages. Analysis time depends on the materials and the hardware.
Does every detected piece of data have to be removed?
No. The scope is determined by the user and the project configuration. Information needed for a particular purpose can remain visible, while specified data or form fields can be redacted. The final document needs to be assessed according to the purpose of disclosure.
Are you preparing a large collection of documents and want to see how its anonymization can be organized? Contact us. During a presentation, we will demonstrate projects, rules, and result verification, and discuss your organization’s needs.
