Nakato is a privacy layer that swaps sensitive data for contextually faithful semantic twins. We let you use restricted data with LLMs, keeping it safe and useful.
Synthetic example. No real customer data appears on this page.
See the swap in four sectorsHow it works
Semantic
Nakato reads the whole document and finds every sensitive detail. This includes indirectly identifying information that pattern matching misses. Why PII detection needs context.
Pseudonymisation
Every sensitive value is swapped for a semantic twin: fictional, but contextually faithful. Same shape, different person.
Reversible
When the answer comes back, the twin is reversed back to the original data. Every substitution is logged and auditable. Real data never left your environment.
| Type | Original | State |
|---|---|---|
| person | Margaret Ellis | found |
| date of birth | 1961-04-12 | found |
| account number | 20-45-11 88451203 | found |
| address | 14 Rowan Grove, Sheffield | found |
4 values found, read for meaning rather than matched.
RSP is our data transformation pipeline. It converts sensitive details into realistic twins at the moment of query, so the model gets the context it needs to work. Your data stays put; Nakato runs inside your VPC or on-premises.
The demo
One layer for every sector. The record on the left stays in your environment, the model only sees the record on the right.
Credit decisions with context
| Field | Original stays with you | Twin what the model sees | Preserved |
|---|---|---|---|
| Customer | Margaret Ellis | Eleanor Grant | Preserved — gender, name origin |
| Date of birth | 1961-04-12 | 1959-11-03 | Preserved — age band |
| Sort code | 20-45-11 | 31-08-64 | Preserved — valid format |
| Balance | £18,240.55 | £17,905.12 | Preserved — within 2% |
Credit decisions with context. Stripping demographics strips the vulnerability signals credit analysis depends on most. The significance of each detail is preserved, meaning your AI tool can reason properly.
Where the two columns match is where details essential to the task are kept. RSP keeps what is needed for the task's reasoning but makes it unidentifiable.
Benchmarking
The field is the benchmark. Each point stands for the same number of test cases. Each mark reaches as far as that tool's figure.
The field is the benchmark. Each point stands for the same number of test cases.
The field is the benchmark. Each point stands for the same number of test cases.
The field is the benchmark. Each point stands for the same number of test cases. Each mark reaches as far as that tool's figure.
Summarisation accuracy on the same cases. A longer bar is better.
Leak rate under the same attack. A shorter bar is better.
Shown alone. A comparison is drawn only where the same benchmark produced one.
Checked across 500,000 financial-services test cases. The full report, with its methodology, is available on request.
How each figure was measuredNakato is built to regulatory requirements for pseudonymised data. A compliance team can see how it meets the rules before anyone asks.
Your real data never leaves your environment.
The model only ever receives the twin, and the mapping remains in your infrastructure. Nothing about your existing data classification changes.
Designed to the regulator's test.
There is no reasonable means of tracing a twin back to a person, which is the test GDPR applies when deciding whether data identifies someone. What teams need for approval.
Every substitution is auditable.
Every swap is mapped (what was replaced, with what, and when) so a reviewer can retrace a decision without ever seeing the original data.
Built against the EU AI Act, DORA, FCA and PRA expectations, GDPR and SOC 2 from the start.
Security signs off on where the data goes. Whoever runs the AI programme wants to know how it works. These are the questions we hear most often.
Yes. Nakato is model-agnostic.
It works with any LLM API, so you keep the model you have chosen for your tool. Whichever you use, it only ever sees your data's twin.
The twin preserves the context needed.
A twin changes what identifies a person but keeps what matters to the task: context like an age band, a format, a balance. That is the highest of the five levels of pseudonymisation. The answer comes back to the user with the originals restored.
Tens of milliseconds.
Detection and twinning happen at the moment of each query. There is no batch job to wait for.
Another question? More questions answered. Or write to hello@nakato.ai and we will answer it.
Join us
Creating semantic twins is not a solved problem. It requires a unique mix of linguistics, semantic ontology and computer science. If you want to work with us to keep developing this solution, talk to us.
Semantic ontology meets computer science. Two disciplines, one twin that keeps its meaning.
The problem changes shape with every domain. What preservation means for medical data looks nothing like insurance.
The work does not repeat. Small team, hard problems, and no template to follow.
We are defining how this is done.