Platform | Pseudia

Product Tour

A real platform, not a mockup.

Every screen below is the actual Pseudia console. Take a look at how a team configures and runs pseudonymisation day to day.

Workspaces

Organise your teams

Create workspaces for each team or department (BI, analytics, engineering) and let each one configure its own data sources and environments independently.

Screenshot of the Pseudia Workspace tab, showing a list of team workspaces

Configuration

Define what gets pseudonymised

Link a configuration to a data source and pick exactly which fields must be pseudonymised, down to the individual column.

Screenshot of the Pseudia Configuration tab, showing fields mapped to a data source

Playground

Preview on sample records

Test a configuration on a handful of sample records before connecting a real source, perfect for quick checks or configurations that have no source yet.

Screenshot of the Pseudia Playground tab, previewing pseudonymisation on sample records

Text Pseudonymisation

Mask sensitive free text

Detect and redact sensitive information directly inside unstructured text, such as emails, notes, or support tickets.

Screenshot of the Pseudia Text Pseudonymisation tab, masking sensitive text

Worker

Run it all at scale

Execute your pseudonymisation configuration as a batch job over an entire data source, monitored from a single dashboard.

Screenshot of the Pseudia Worker tab, showing a batch pseudonymisation job

Features

Concrete, not promises.

01

Keys separated from data

An architecture where keys never co-reside with pseudonyms. A leak exposes pseudonyms, not people.

CISO

02

Test environments in minutes

Generate realistic datasets on demand while preserving structure, distribution, and relational consistency.

Innovation

03

AI gateway for LLMs

Pseudonymise prompts, documents and RAG sources before they reach external LLMs. Your teams keep their preferred AI tools.

AI

04

Policy-based exportable audits

Configure, trace, and export. Every processing activity is documented and auditable. Prove compliance, don’t just claim it.

DPO

05

Custom BYO functions

Plug in your own algorithms via API when our native functions don’t cover your specific sectoral needs.

Flexibility

06

100% European & Sovereign

Designed and hosted in Europe. Zero data transfers outside the EU by default. On-premises deployment available.

EU Sovereign

Understand

Not all approaches
are equal.

Before choosing, you need to know what is really being compared.

Common practice

Naïve name removal

Only the name is stripped. Everything else (IBAN, age, postal code) still circulates freely. By cross-referencing just a few fields, 87 % of individuals can be re-identified. (MIT, Sweeney)

⚠ Not true anonymisation. GDPR obligations remain in full.

Original data

Name Marie Dupont

IBAN BE68 5390 0754 7034

Age 34

Postal code 4000

→

Result

Name ████████ deleted

IBAN BE68 5390 0754 7034 exposed

Age 34 exposed

Postal code 4000 exposed

Rigorous anonymisation

k-anonymity · differential privacy

Mathematically grounded. Each individual becomes indistinguishable from at least k others, so re-identification by cross-referencing becomes very unlikely.

ℹ Can take data outside GDPR scope when done right. However, it degrades data precision (ranges, regions) and demands significant implementation effort.

Original data

Name Marie Dupont

IBAN BE68 5390 0754 7034

Age 34

Postal code 4000

→

Result

Name ████████ redacted

IBAN ████████ redacted

Age [30–35] generalised

Postal code 4*** generalised

The AI-generated path

Synthetic data

A model generates a fictional dataset statistically close to reality. No real individual appears in the output.

⚠ The model has ingested your real data. Rare profiles can leak (membership inference). Cross-table consistency is often incomplete.

Original data

Name Marie Dupont

IBAN BE68 5390 0754 7034

Age 34

Postal code 4000

→

Result

Name Laura Bernard fictional

IBAN BE12 3456 7890 0000 fictional

Age 33 approx.

Postal code 4020 approx.

The Pseudia approach

Pseudonymisation

Real identifiers are replaced by consistent tokens or realistic fake values. Marie Dupont → psd_k8x2m. Indirect identifiers such as postcodes can be made less precise. Structure, distribution and relationships remain intact. A separate key allows controlled reversal, logged and audited.

✓ Reversible, realistic, controlled. A safeguard recognised by GDPR art. 4.5 · 25 · 32: the data remains personal data for whoever holds the key, with drastically reduced exposure.

Original data

Name Marie Dupont

IBAN BE68 5390 0754 7034

Age 34

Postal code 4000

→

Result

Name psd_k8x2m token

IBAN BE71 0961 2345 6769 realistic fake

Age 34 intact

Postal code 40** generalised

Name removal

True anonymisation

Synthetic

Pseudonymisation

Real person protected

✗ No

✓ Yes

~ Partial

~ Yes, residual risk¹

Personal data (GDPR)

✓ Yes

✗ No

✓ Still yes¹

Precise & usable data

~ Degraded

~ Approx.

✓ Intact

Cross-table consistency

~ Partial

✓ Complete

Reversible with authorisation

✗ No

✓ Yes, separate key

Implementation effort

✓ Minimal

✗ High

~ Medium

✓ Low with Pseudia

Dynamic data support

✓ Yes

✗ Poor

✗ Batch only

✓ Excellent

Recommended scope

~ Limited

~ Targeted outputs

~ Dev / sharing

✓ Everything, continuously

¹ Pseudonymised data technically remains personal data as long as the re-identification key exists. GDPR applies. Pseudia drastically reduces real exposure risk, but does not replace the associated legal obligations.

Architecture

How Pseudia processes your data

Three steps. Secure. Sovereign. Reversible.

① Source

Sensitive Data

Your production databases, APIs, customer files, or logs containing PII (Personally Identifiable Information).

PostgreSQL MySQL CSV / JSON Parquet S3

ingest

② Pseudia Engine

Automated Pseudonymization

Identifiers are replaced by consistent, realistic tokens. Statistical distribution is preserved, and re-identification is only possible via a separate key.

TokenizationMaskingBYO functionsPolicy-driven

output

③ Target

Secured Data

Pseudonymized data is delivered to your destination: test environments, AI pipelines, partners, or analytics.

Test envLLM pipelinesAnalyticsPartners

🔑

Sovereign Decryption Key
Stored separately from pseudonymized data. Only authorized users can reverse the process. If a data breach occurs, captured pseudonyms do not reveal who people are without the key.

Free resource

Something worth printing, for once.

A practical, two-page checklist for security and data teams, coming soon. Leave your email and we'll send it to you as soon as it's available.

The Data Pseudonymisation Readiness Checklist

12 concrete checks to run before plugging real data into test, AI or analytics pipelines, covering data mapping, technique selection, key management, and how to make your process provable to a DPO.

Know exactly where your personal data lives, and where it leaks
Choose the right technique: masking, tokenisation or synthetic data
Make your pseudonymisation process provable, not just believable

PDF · 2 pages · EN & FR · free · coming soon

📄 PDF

Enter your work email and we'll send you the PDF as soon as it's released. No spam, unsubscribe anytime.

One platform to turn restricted data into usable data.

A real platform, not a mockup.

Organise your teams

Define what gets pseudonymised

Preview on sample records

Mask sensitive free text

Run it all at scale

Concrete, not promises.

Keys separated from data

Test environments in minutes

AI gateway for LLMs

Policy-based exportable audits

Custom BYO functions

100% European & Sovereign

Not all approaches
are equal.

Naïve name removal

k-anonymity · differential privacy

Synthetic data

Pseudonymisation

How Pseudia processes your data

Something worth printing, for once.

The next leak will not wait
for your next committee.

A real platform, not a mockup.

Organise your teams

Define what gets pseudonymised

Preview on sample records

Mask sensitive free text

Run it all at scale

Concrete, not promises.

Keys separated from data

Test environments in minutes

AI gateway for LLMs

Policy-based exportable audits

Custom BYO functions

100% European & Sovereign

Not all approachesare equal.

Naïve name removal

k-anonymity · differential privacy

Synthetic data

Pseudonymisation

How Pseudia processes your data

Something worth printing, for once.

The next leak will not waitfor your next committee.

Not all approaches
are equal.

The next leak will not wait
for your next committee.