Science-backed anonymisation
Works 100% offline

Anonymize Text.
Retain Data.

Remove personal data from interviews, documents, and logs — directly on your device. For Mac, Windows, Linux, and mobile, or at scale via cloud API.

Works with English, German, Dutch, French, Spanish, Italian, and many more languages

Developed by Prof. Dr. Bennett Kleinberg & jocapps® GmbH — validated in peer-reviewed studies

See benchmark reports & evidence

Trusted by

Developed for Smart Privacy

No coding required An intuitive app anyone on your team can use — no technical setup
Your data stays with you Everything is processed locally — nothing is ever uploaded
Understands context AI-based detection recognises personal data from context — not just keyword lists
Proven in studies Validated in tests where humans tried — and failed — to re-identify people

Core Principles

Scientific Foundation

Built on the open Textwash™ research project — transparent, auditable, developed by academic researchers

Contextual Privacy

Anonymises based on linguistic context, not just keyword matches

Local-First Architecture

Made for sensitive data: the desktop app needs no internet connection at all

ISO 9001 & ISO 27001 Certified Development Company

Textwash™ in action

Anonymisation examples

From a simple sentence to messy real-world text: Textwash™ recognises name variants, initials, and context — and maps them consistently to the same placeholder. Switch between examples and languages:

Examples
Languages
Input Text

"Lisa works at Apple and lives at 123 Baker St."

Textwash™ Pro
Anonymised Output

"PERSON_1 works at ORG_1 and lives at ADDRESS_1."

Many ways to use Textwash™

Use it as a fully local end-to-end tool — or as a local privacy layer in front of cloud AI services. Either way, the anonymisation itself always happens on your device.

Classic: anonymise & work fully local

For researchers, clinicians, and analysts: anonymise interviews, notes, and documents directly on your device — then analyse, archive, or share the anonymised copies without privacy risk.

🔒 Your device — nothing is uploaded

📄 Original text with personal data
Textwash™ anonymises — 100% locally, even offline
✅ Anonymised text
📊 Analyse · share · archive

AI proxy: pre-clean text for OpenAI, Claude & Co.

Want to use cloud LLMs on sensitive text? Textwash™ anonymises everything locally first — the cloud service only ever receives the already-anonymised version.

🔒 Your device — anonymisation happens here

📄 Original text with personal data
Textwash™ anonymises — 100% locally
↓  only the anonymised text is sent  ↓

☁️ Cloud — optional next step

🤖 OpenAI · Claude · any LLM — receives anonymised text only

🚫 The anonymisation is never done by OpenAI, Claude & Co. — Textwash™ does it on your device, and the original, identifiable text never leaves your machine.

...and many more!

Product family

One anonymisation engine, four ways to use it: desktop app, mobile, cloud, or API

Desktop app · 100% offline

A look inside Textwash™ Pro on Mac, Windows, and Linux:

Before / after anonymisation with codebook & statistics
Configurable entity types & detection settings
Optional rule builder for custom exceptions & patterns
Desktop & mobile app

Textwash™ Pro

Mac · Windows · Linux · iOS · Android

Import text, anonymise, export — entirely on your device. Nothing is sent to external servers.

  • Most advanced AI models
  • English, German, Dutch, French, Spanish, Italian & many more
  • 100% local & offline — air-gap ready
  • Whitelist & blacklist rules
  • Advanced reporting: codebook & statistics
  • Super-fast: approx. 3 documents per second
Offline by default · GUI-based
API & integrations

Textwash™ Pro API

Cloud-based processing · Ready for n8n, Make, & Zapier

Build anonymisation into your own systems and pipelines — from web apps to automations with n8n, Make, and Zapier.

  • Most advanced AI models
  • Multiple languages
  • REST API — ready for n8n, Make & Zapier
  • Whitelist & blacklist rules
  • Scales to large document volumes
REST API · Integrations
Cloud workspace

Textwash™ Pro Cloud

Browser-based batch processing

Run anonymisation jobs in the browser — hosted by us or in your organisation's own cloud.

  • Most advanced AI models
  • Multiple languages
  • Browser-based batch processing
  • Team projects, dashboards & result logs
  • Hosted by us or in your own cloud
Hosting in your cloud or servers by us · Team-ready
Textwash™ Free (open source) vs. Textwash™ Pro — feature comparison

Textwash™ Pro builds on the open-source original and extends it with the most advanced models, more languages, and a complete workflow around them.

Textwash™ Free · open source
Textwash™ Pro
AI models
Basic models
Most advanced models, continuously updated
Languages
English + Dutch
English, German, Dutch, French, Spanish, Italian & many more
Interface
Script only (command line)
Desktop & mobile apps, cloud & API
100% local & offline
✓
✓ incl. air-gapped environments
Processing speed
Depends on command-line skills and CPU
Super-fast: approx. 3 documents per second; parallelised via CPU + GPU (see reports)
Whitelist — keep chosen terms untouched
—
✓ incl. visual rule builder
Blacklist — always anonymise custom terms & patterns
—
✓ incl. visual rule builder
Configurable entity types
Fixed set
16 selectable entity types + detection sensitivity
Create your own schemas
—
✓
Custom label mapping
—
✓
Anonymisation risk score
—
✓
Risk sensitivity filtering
—
✓
Interactive before/after view
—
✓
Advanced reporting
—
✓ Codebook, statistics & coverage reports
File formats: PDF, DOCX & XLSX
Plain text only
✓ plus TXT & CSV
Batch processing
Manual scripting
✓ Files & whole folders
Built-in tutorials
—
✓
Support & updates
Community (GPL-3.0)
Professional support & SLA options

Textwash™ Free — the open-source original (GPL-3.0): script-based, no GUI, for technical users who want the code itself.

Source code & paper

Typical use cases

Built for real-world anonymisation in research, industry, and the public sector:

Working with text that contains personal data? Not sure if it fits? Ask us at textwash-pro@jocapps.eu

GDPR- & HIPAA-compliant data anonymisation

Anonymise support logs, email archives, contact forms, and CRM notes before they are stored or shared

Archiving & records retention

Reduce privacy risk in legacy case files, document collections, and internal knowledge bases — without losing their value

Open Science & data sharing

Share survey responses, interview transcripts, and qualitative research data while protecting participants

Legal, Health, & Social services

Remove identifiers from clinical notes, legal case summaries, and social work documentation

User research & UX feedback

Share user interviews, usability tests, and support tickets safely across teams and partners

Logs & monitoring data

Strip personal data from application logs, chat histories, and audit trails before central storage or analysis

Safe inputs for AI & LLM workflows

Anonymise prompts, tickets, and free-text inputs before they reach internal or external AI systems — a privacy layer in front of your LLM pipeline

Custom institutional workflows

We design end-to-end anonymisation workflows that fit your governance, legal, and research quality requirements.

Governance alignment

Policy mapping, retention rules, and approval checkpoints

Controlled processing

Role-based access, review loops, and privacy controls

Audit readiness

Documented procedures and repeatable validation evidence

For institutional rollouts, integration planning, or compliance questions, contact textwash-pro@jocapps.eu

Optional services

Textwash™ Pro works out of the box. If you want support, we offer implementation and consultancy services for research teams, companies, and the public sector.

Advisory and implementation support

Operational design, integration planning, and quality assurance for sensitive text workflows — tailored to your setup.

  • Workflow assessment for privacy, compliance, and data utility
  • Quality review strategies for high-impact datasets
  • Anonymisation layers for AI and LLM pipelines
  • Integration across on-prem and cloud environments

Phase 1

Discovery

Review your data landscape, map risks, and define the target workflow

Phase 2

Pilot

Onboard datasets, calibrate entity types, and set up quality review

Phase 3

Integration

Connect systems and APIs, with runbooks and monitoring in place

Phase 4

Governance

Audit evidence, policies, training, and continuous improvement

Optional services are available for SMEs, enterprise teams, universities, healthcare, and public sector

Built for serious data protection work

Four principles guide how Textwash™ Pro is built:

1. Transparent evaluation

The approach has been tested empirically — including intruder tests where humans tried to re-identify people in anonymised documents.

2. Data never leave your system

No uploads, no remote APIs. Disconnect from the internet and keep anonymising.

3. Open foundations

Built on the open-source Textwash™ research project — inspectable, testable, and extendable by anyone.

4. Anonymisation that understands language

Personal information depends on context. A machine learning model reads each phrase in context instead of matching against fixed word lists.

Comparing anonymisation tools?

Whatever tool you choose, ask the provider for two things: an empirical evaluation of what it can and cannot do, and a clear reason why your data would need to leave your systems at all.

If that transparency is missing, treat risk claims with caution.

Questions about evaluation details? textwash-pro@jocapps.eu

Data Protection Laws (GDPR & HIPAA)

Compliance by design, not as an add-on

Textwash™ Pro is 100% ready for current EU privacy requirements and is compliant with GDPR and HIPAA — supporting data minimisation, purpose limitation, and privacy by design/default (Articles 5 and 25 GDPR).

Developed by an ISO 9001 & ISO 27001 certified company.

European AI Sovereignty

Full local deployment on Windows, Linux, and macOS

The Windows/Linux/macOS app runs fully local, offline, and air-gapped. No data leaves the client environment, and no external APIs are required.

Frequently asked questions

Deployment models, support levels, and governance requirements

Is Textwash™ Pro usable without optional services?

Yes. The product is fully usable on its own, and services are optional

Do you provide SLA options?

Yes. We can define service levels, support windows, response targets, and escalation paths for qualifying organisations

Is Textwash™ Pro suitable for public sector programmes?

Yes. We support public sector, research, healthcare, and regulated environments with governance-aligned implementation plans

Can on-premise and cloud setups be combined?

Yes. Hybrid architectures can combine local processing with API or cloud components, depending on policy and risk constraints

How do you support audits and compliance reviews?

We provide documentation inputs, quality checkpoints, and implementation evidence to help internal governance and external audits

Who should contact you for enterprise or institutional rollout?

Programme managers, data protection teams, and technical leads can contact us at textwash-pro@jocapps.eu to discuss fit and rollout options

How does Textwash™ Pro support GDPR compliance in practice?

Textwash™ Pro is aligned with GDPR principles including data minimisation, purpose limitation, and privacy by design/default (Articles 5 and 25), and supports compliance-focused workflows for sensitive text handling

Can Textwash™ Pro be used in sovereign or air-gapped AI environments?

Yes. The Windows/Linux/macOS deployment runs fully local and offline, so no data leaves the client environment and no external API connectivity is required

How it works

Anonymising a text collection takes four steps — no command line, no technical setup:

  1. Open Textwash™ Pro on Mac, Windows, Linux, iOS, or Android
  2. Import files or whole folders
  3. Pick the language and an output folder
  4. Run — anonymised copies are saved to the chosen folder

Works equally well for a single document or large text collections.

Need a walkthrough?

If you would like a short demo or have specific questions about your use case, we are happy to help.

Examples & sample data

Want to see real results first? The open-source Textwash™ project includes detailed example texts together with their anonymised counterparts — a good starting point for your own evaluation.

Browse Textwash™ Free on GitHub

You decide what gets anonymised

Choose exactly which entity types to anonymise — and keep everything else intact.

Align anonymisation with legal and methodological requirements while preserving as much useful information as possible.

PRONOUNS PHONE NUMBER EMAIL ADDRESS NUMERICS MONTHS DATE PERSON LOCATION OCCUPATION TITLE AGE CULTURAL IDENTITY TIME ADDRESS ORGANISATION OTHER IDENTIFIABLE ATTRIBUTE

Research & evidence

We don't just claim strong privacy — independent, peer-reviewed research backs it up. Textwash™ has been benchmarked against multiple anonymisation approaches and evaluated in published studies, giving procurement and governance teams evidence they can cite.

Technical reports

Privacy in benchmark comparisons

Independent benchmarks rank Textwash™ as highly competitive and dependable on privacy performance

Honest trade-offs, not single-metric claims

Published studies weigh privacy against data utility and cost — the basis for realistic deployment decisions

Peer-reviewed evidence base

Journals and conference proceedings you can reference in procurement, governance, and documentation

• arXiv (2026): arxiv.org/pdf/2602.12806

• Procedia Computer Science (2025): sciencedirect.com/science/article/pii/S1877050925008518

• arXiv (2024): arxiv.org/abs/2411.05978

• Nature Scientific Reports (2023): nature.com/articles/s41598-023-42977-3

• ACM Digital Library (2023): dl.acm.org/doi/abs/10.1145/3576050.3576070

• arXiv (2021): arxiv.org/abs/2103.09263

• and many more

Benchmark comparison from the technical report
Click to open benchmark figure on arXiv

For research collaborations, interoperability discussions, or evaluation questions, contact textwash-pro@jocapps.eu

Become a Textwash™ Pro partner

Looking for an anonymisation partner for your product, organization, or research project? Let’s discuss integrations, pilots, and custom deployment options.

Email us

Who developed Textwash™ Pro?

Textwash™ Pro is developed and distributed by Prof. Dr. Bennett Kleinberg & jocapps® GmbH and is based on Textwash™ (github.com/ben-aaron188/textwash) under the GNU General Public License v3.0. The original Textwash™ project was developed by Dr. Maximilian Mozes and Prof. Dr. Bennett Kleinberg.

Textwash™ Pro extends this foundation with a multi-platform GUI, deployment options, and additional tooling while preserving the open, research-driven ethos of the original project.

Paper: Kleinberg, B., Davies, T., & Mozes, M. (2022). Textwash, automated open-source text anonymisation. arXiv:2208.13081.

Questions? Mail us!
textwash-pro@jocapps.eu