Local protection for business AI use
AVGfilter. Work with AI. Protect your data locally.
The browser extension protects personal data on your own device before you send text to AI. Doc Scanner masks documents and photos on your computer. No upload to a filtering service is needed.
- Local protection without uploading your original data to a filtering service
- Blocks sending when protection cannot be guaranteed
- Adds control without imposing a new workflow
AVGfilter
Preparing a service quote
On your own device
The browser extension detects and replaces personal data locally before sending.
Only protected text goes to AI
Detected data is replaced with tokens. The remaining text goes to the AI service you use.
Original files stay with you
Doc Scanner processes documents and photos on your computer, even without internet.
Why it matters
AI accelerates work. Data protection must keep pace.
Names, addresses, case files and financial details easily end up in a prompt. AVGfilter protects data at the point of entry, not through an audit afterwards.
Invisible use
AI is used throughout the organisation. Central policy cannot see what is actually entered.
Irreversible sending
Once sent, sensitive information sits outside your direct technical control.
Ongoing responsibility
Your organisation remains responsible for careful processing of personal data.
Broad protection
More than names and phone numbers.
AVGfilter detects a broad set of data and can additionally protect organisation-specific terms.
Works inside
How it works
Protect locally. Then send.
- 01
Detect
The browser extension checks input locally on the employee’s device.
- 02
Replace safely
Detected personal data is replaced locally with tokens. Only protected text goes to AI.
- 03
Restore locally
The browser extension restores tokens locally in the response. The link to original data stays on your device.
For organisations
From individual protection to demonstrable control.
AVGfilter combines immediate employee protection with the manageability organisations need.
Policies and custom terms
Tune detection and protected terminology to your organisation.
One approach across teams
Apply the same safety layer across multiple AI tools and employees.
Insight and accountability
Support reviews, evaluation and demonstrably careful AI use.
Built for daily work
For work that needs context, but not identity.
Evaluate it yourself
See exactly what the AI model does and does not receive.
Use the sample to see which data is replaced. This online demo uses a filtering service and is not local: do not enter real confidential data. The browser extension and Doc Scanner process locally.
Privacy filter, live
This online demo sends your input to a filtering service. Use the sample, not real confidential data. The browser extension and Doc Scanner work locally.
- Received
- Tokenized
- AI processing
- Restored
- Completed
Input
text from the employeeDetected personal data
Run the filter to see which data gets replaced.
Tokenized version
sent to the modelNo tokenization performed yet.
What the model receives
literal payloadNothing sent yet.
Detected data is replaced with tokens. The remaining text stays visible to the AI service.
Answer restored
readable for the employeeThe answer is made readable again after processing.
Keep your data protection local.
See it work live or discuss a pilot for your organisation.