Document anonymization
Anonymize documents: placeholders instead of black bars
skai reads scanned files as PDF page by page and replaces names, addresses, dates of birth and other personal data with placeholders such as [First name] or [Address]. You review uncertain passages on the scan. The result downloads as a PDF in which the replaced data is removed from the image.
The problem
Why redacting by hand falls short
-
Page by page
Redacting a file with a marker or a PDF tool takes time, and it starts over for every new recipient.
-
Easy to miss
Handwritten entries, form fields, signatures and pasted-in photos slip through when you leaf through a file.
-
Only covered
A black bar over the text does not remove it in every PDF. Underneath it often stays readable.
How it works
Five stages per file
- 1
Hand over
Upload the scanned PDF and give it a name.
- 2
Read
skai reads every page and marks passages it could only decipher with doubt.
- 3
Replace
The run sets the placeholders and flags data where it is unclear whether it must be replaced.
- 4
Review
You go through the uncertain passages on the scan, correct read text and decide every open item.
- 5
Close
The reviewed document becomes immutable. skai creates the anonymized version as a PDF to download.
One rule set for every file
What gets replaced and what stays readable
Replaced
Every recognised item gets the placeholder of its kind. Every private person’s name is replaced, whoever it belongs to.
- [First name]
- [Last name]
- [Date of birth]
- [Place of birth]
- [Address]
- [Place of residence]
- [Phone number]
- [E-mail address]
- [File number]
- [Insurance number]
- [Tax ID]
- [ID number]
- [Bank details]
- [Licence plate]
- [Employer]
- [Signature]
- [Religion]
- [Date]
- [Customer number]
- [Image]
Stays readable
So the file stays understandable, skai deliberately keeps these.
- Authority and departments
- Official roles
- Legal bases
- Legal entities
- Procedural dates
- Amounts
- Nationality
- Marital status
In detail
Built for real files
-
Placeholders, not pseudonyms
A placeholder only names the kind of data. “Anna Weber” does not become a “Person A” you could recognise across the file.
-
Handwriting and forms
Handwriting is treated like print. In forms skai judges a field value together with its label and replaces only the value.
-
Images and signatures
Images are covered as a whole, signatures become [Signature]. In the final PDF the data is removed, not just covered.
-
Only you see your files
Submitted documents do not go into a knowledge base. Other members of your organisation, administrators included, cannot see them.
Ask questions without revealing personal data
For questions to the AI there is the pseudonymized chat: skai replaces name, address and numbers with placeholders before the question reaches the AI model.
Video: The pseudonymized chatQuestions about anonymization
Which files can I submit?
Scanned documents as PDF. Other formats such as Word are not accepted; export or scan the file as a PDF first.
Which data gets replaced?
Among others first and last name, date and place of birth, address, phone number, e-mail address, file number, insurance, tax and ID numbers, bank details, licence plate, employer, signature and images. If a kind is missing, you create your own placeholder; skai then searches the whole file for further occurrences.
Why is the name of the authority still in the document?
That is intended. Authority, departments, official roles, legal bases, legal entities, procedural dates and amounts stay readable so the file remains understandable.
Is the data removed from the final PDF or just covered?
Removed. In every box with replaced data the scan is painted over and the anonymized text laid on top; covered images are black.
Who can see the documents I submit?
Only you. Other members of your organisation, administrators included, cannot see your submitted documents.
My run stopped. Do I pay twice?
No. If the run can be resumed, it continues from the last finished page; only newly read pages are charged to your credit.
See it on one of your files
In a no-obligation demo we show the anonymization and discuss which documents you deal with.
Book a demo
Talk to us
With skai we want AI to arrive in everyday work: answers from your own knowledge, with sources, without anyone having to learn prompting. Write to me, call me or book a slot directly.
Stefan Henke
Management