Runs on your own machines

Let your team use AI
without handing over your data.

OmitPII sits between your people and the AI tools they already use. It hides names, account numbers and other sensitive details before anything leaves the building, then restores them in the reply you get back.

Get started See how it works Download the datasheet (PDF)
Built to help you meet EU AI Act Art. 12, UK GDPR and Jersey DPL 2018 record-keeping duties.
For your team

Day-to-day AI, minus the risk

People can use ChatGPT, Claude, Copilot or whatever helps them get work done, without sensitive data going along for the ride.

For your systems

Agents that stay in bounds

Let automated tools and agents handle work on your behalf while the personal data they touch stays masked the whole way through.

Who we're talking to:
UK & Channel Islands banks Wealth and trust firms Healthcare providers Legal and professional services
01 / THE PROBLEM

The fines are the easy part to picture

Everyone wants the time AI saves, but nobody wants to be the company that leaked client data into a chatbot. Most redaction tools want you to upload the document to their cloud first, which rather defeats the point, and doing it by hand doesn't scale past a handful of files.

SOURCE: FINANCIAL REGULATORS & ICO
UK GDPR
ICO Enforcement
£17.5M or 4% of Global Turnover

Applies to unlawful processing of data or failing to comply with DPA 2018 accountability requirements. Maximum penalty applies.

EU AI Act
EU Governance Board
€35M or 7% of Global Turnover

Serious violations regarding data governance and training datasets. Full high-risk system compliance applies from 2 August 2026.

Jersey DPL 2018
JOIC Enforcement
£10M or 10% of Global Turnover

Mirrors GDPR principles but with heightened accountability guidelines for local financial and trust operations.

Corporate Exposure Calculator Live Risk Projection
Projected Maximum Statutory Fine Caps:
UK GDPR
£17,500,000
EU AI Act
€35,000,000
Jersey DPL
£10,000,000
* Calculations based on legal maximums: UK GDPR (higher of £17.5M / 4%), EU AI Act (higher of €35M / 7%), Jersey DPL (higher of £10M / 10%).
02 / HOW IT WORKS

How It Works

OmitPII runs on your own machines. It swaps sensitive details for stand-in tokens before a document is sent anywhere, and swaps the real values back into the reply once it returns. The originals never leave your network.

Light on IT It works alongside the AI tools your team already has, so there's nothing to rip out or replace.
01 / Local
Hide the details
Names, emails, account numbers and the like are swapped for tokens on your own machine.
02 / Transit
Send the clean version
Only the tokenized text goes out to the model you've chosen (Gemini, OpenAI, Claude), using your own keys.
03 / Process
Let the AI work
The model answers the question or runs the task without ever seeing a real identifier.
04 / Restore
Put the details back
When the reply comes back, the real values are dropped back in, locally, before you read it.
See it on a sample document Status: Active
What you drop in Stays local
Loading simulation data...
What actually gets sent Leaves the network
Loading simulation data...
Current state: Reading the document

What it looks for

Type of data Examples What happens to it
Personal details Names, personal emails, phone numbers, IP addresses Swapped for a token
Financial Card numbers, IBAN/BIC, bank account numbers Removed
National identity UK National Insurance, EU social security, US SSN Swapped for a token
Your own patterns Account references, internal keys, anything you can describe with a rule Masked

How it's built

Nothing phones home

No usage tracking, no calls back to us. OmitPII does its job inside your network and stays there, which is what sovereign and regulated environments need.

Fits how you already work

Run it as a lightweight app on Windows desktops, or as an on-premise gateway in front of your databases and internal APIs.

Encrypted and kept local

The map between tokens and the real values is encrypted and stored on your own systems. We never hold it, and it never reaches us.

03 / PRICING

Plans

One flat price per plan, so you can budget for it. No per-seat charges and no surprise bills tied to how much your team uses AI.

Pick by size Choose the plan that matches how many machines you need to cover.
Pilot
£200 / month30-day trial
  • Up to 5 machines
  • Standard data filters
  • Runs locally, nothing stored
  • Community support
Start a pilot
Team
£3,500 / yearBilled annually
  • Up to 25 machines
  • Built-in personal data filters
  • Local audit log
  • Email support, business hours
Choose Team
Most popular
Business
£12,000 / yearBilled annually
  • Up to 150 machines
  • Your own custom filters
  • Help getting set up
  • Compliance certificates
  • Priority support (8hr response)
Choose Business
Enterprise
£25,000 / year+ £5,000 setup
  • Unlimited machines
  • Server-side gateway
  • Connects to your databases
  • Covers automated agents too
  • 24/7 phone support (2hr critical)
Talk to us

Want to try it across a whole organisation first?

We'll run a 30-day trial with you and help with the setup, the on-premise gateway, and any filters specific to your data, so you can see it working before you commit.

Set up a trial
04 / GET STARTED

A few quick questions

Tell us a bit about how your team uses AI today. It takes a couple of minutes, and we'll come back with where the gaps are and how OmitPII would fit.

Your answers stay between us.
Step 1 of 10: Your role

How would you describe your company's current stance on employees using AI tools (like ChatGPT, Claude, Copilot)?

How often do you or your team handle documents containing Personally Identifiable Information (PII) or sensitive company data?

Not At All All The Time

Have you ever wanted to use an AI tool to summarize or analyze a document, but stopped because it contained sensitive names, addresses, or financials?

With upcoming regulations (like the EU AI Act) and stricter GDPR enforcement, how confident are you that your company could survive a surprise audit regarding what data has been fed into LLMs?

Terrified Extremely Confident

Imagine a software that sits between you and your AI. You drop a document into it, and it instantly hides all sensitive information before sending it to the AI. When the AI replies, the app instantly puts the real data back in the document(s) for you to read.

How valuable would this be to your daily workflow?

Useless Game Changer

What is the most appealing part of that concept to you?