Pseudonymised before the text reaches the AI
Talk to the AI, with personal data pseudonymised automatically before it gets there.
What is detected (names, tax ids, ΑΜΚΑ social security numbers, IBANs) becomes [NAME_1] before it reaches the AI provider, and comes back real in the answer, on your side. Inside the filterit chat, or with the agent under the popular AI chats you already use.
Start for free Try it with your own text
Solutions by sector
Built for the way your office works
The same tools, with examples, data categories and guides for each sector.
Law firms
Pleadings, contracts and correspondence without the client leaving the office.
Accounting firms
Payroll, filings and client email without tax ids leaving.
HR teams
CVs, employee records and payroll with the names, tax ids and ΑΜΚΑ (social security numbers) that are detected masked.
Clinics and practices
Reports and referrals with the names and ΑΜΚΑ that are detected masked.
Public sector
Applications, decisions and briefs with the citizen details protected.
Developers and product teams
API, SDKs, MCP and IDE plugins: code, data and secrets are masked before they leave for the AI assistant.
Two ways to use it
AI chat, or the agent on your PC. Same filter.
AI chat inside filterit
Chat with leading models inside filterit. The sensitive details that are detected leave as placeholders, the answer comes back with the values restored, and the audit panel shows what left.
The agent under the AI tools you already use
Keep using the popular AI chats or your IDE assistant. The agent runs on your computer and makes the same replacement locally, before the text leaves for the provider.
Documents, email and scans
Word, PDF, Excel, images and scans with OCR: preview the findings, pseudonymise or redact for good, export to Word or PDF and restore later.
The product
What it does, at a glance
In the filterit chat, in the tools you already use, or on a document. No change to your workflow.
- You write as usual — In the chat, in a popular AI chat or your IDE assistant, or you upload a document.
- Detection — The engine detects names, tax ids, ΑΜΚΑ (social security numbers), IBANs and 30 more categories, with validation and an accuracy we measure and publish.
- Placeholders reach the model — [NAME_1], [AFM_1]: stable within the conversation, so the AI answers correctly without receiving the details that were detected.
- Restore — The answer comes back with the real values, only on your side.
Bilingual language engine
Understands inflections, accents and names in Greek and English, not just text patterns.
Policy and reports for organisations
Mandatory categories per organisation, roles, PII-free reports and an evidence pack for your DPO. Supports the organisation's AI-use policy (GDPR, EU AI Act).
Multiple formats
PDF, Word, Excel, CSV and images, all handled the same way.
What gets detected
34 categories, 11 of them specific to Greek documents
Greek identifiers are validated (check digit, format, context) rather than matched as digit runs, and names are followed across grammatical cases. Tap a category for details.
Greek-specific identifiers
- Tax ID (ΑΦΜ) — Greek tax identification number with modulo-11 check digit validation, not a bare 9-digit match. (check: modulo 11)
- Social security (ΑΜΚΑ) — Greek social-security number with Luhn and birth-date validation. (check: Luhn)
- ID card number — Greek police ID card, old and new formats (letters and digits). (check: format)
- Vehicle plates — Greek registration plates, restricted to the letters Greek plates actually use. (check: format)
- Court case numbers (ΓΑΚ, ΕΑΚ, ΑΒΜ) — Greek court case and docket numbers, context-gated with NLP. (check: context)
- Business registry (ΓΕΜΗ) — General Commercial Registry number, 12 digits, context-gated to avoid clashing with other numbers. (check: context)
- Cadastre code (ΚΑΕΚ) — National Cadastre property code, 12-digit and full 18-digit forms, with prefecture-code validation. (check: prefecture code)
- myDATA and EFKA registries — AADE myDATA MARK and invoice UID, EFKA employer and insured numbers (ΑΜΕ, ΑΜΟΕ). (check: context)
- Property references (ΑΤΑΚ, ΠΕΑ) — E9 property identifier and energy-performance certificate security numbers. (check: context)
- Tiresias, auctions, BIC — Greek credit-bureau references, e-auction codes and BIC/SWIFT codes. (check: context)
- Greek places — Greek gazetteer of cities, municipalities and regions, in every grammatical case.
European and international
- IBAN — Bank accounts of every EU/EEA country with MOD-97 validation. (check: MOD-97)
- EU and UK VAT numbers — VAT numbers of all 27 EU countries and the UK: per-country format check, check digit for Greece, Italy, Poland, Germany and the UK. (check: VIES)
- National IDs (EU, US, UK) — DNI, PESEL, Codice Fiscale, NIR, SSN/ITIN/EIN, National Insurance and more. (check: national checks)
- Passports — Passport numbers gated by context words in the 24 EU languages (automated tests for 4). (check: context)
- Driving licences — Driving-licence numbers, context-gated. (check: context)
- Phone numbers — Landlines and mobiles for the EU, UK, US/Canada and Switzerland, validated per country. (check: libphonenumber)
- Payment cards — Card numbers with Luhn, together with CVV and expiry. (check: Luhn)
- Chassis number (VIN) — 17-character vehicle identification number per ISO 3779, context-gated. (check: ISO 3779)
- Digital signatures — E-signature identifiers: DocuSign envelope, Adobe Sign agreement, SHA-1/SHA-256 certificate thumbprints. (check: format)
- Postal codes — Greek postcodes, US ZIP codes and UK postcodes. (check: context)
People and contact details
- Names across grammatical cases — Greek NER that follows inflection, greeklish spellings and links the forms of the same person.
- Companies and organisations — Company and institution names, Greek and international.
- Email addresses — Email addresses. (check: format)
- Street addresses — Street, number and area, in Greek and Latin-script text.
- Dates — Birth dates and other identifying dates. (check: format)
- Trademarks — Trademark names from a private list of your own clients.
Health and special categories
- Medical data — Blood type, record, insurance and prescription numbers, ICD-10 codes, and a condition written behind a caption or as the reason for a sick leave. A condition inside a free sentence usually stays readable.
- Special categories — Gender, citizenship, religion, marital status and profession written in a captioned form field, and the specialty in staff tables. Inside a free sentence these, like political or trade-union references, usually stay readable.
Code, secrets and systems
- API keys and secrets — API keys, tokens and JWTs, private keys, connection strings, passwords, typed in the placeholder. (check: format)
- Reference codes — Prefixed identifiers such as EMP-, INV-, CUST-, ORD-. (check: format)
- MAC addresses — Network device identifiers. (check: format)
- IP addresses — IPv4 and IPv6. (check: format)
- Technical identifiers — File paths carrying a username and internal hostnames. (check: format)
In scanned PDFs and images, signatures and stamps are covered by visual redaction, with a preview before it is applied.
Integrations
Works where you work
The same detection engine, inside the tools you already use: ChatGPT, Claude and other AI chats in the browser, your IDE's AI assistant, desktop apps through MCP.
- Browser extension — Masks before Enter in the popular AI chats in the browser. Chrome and Edge, manual install.
- Word add-in — Mask and restore inside the document, paragraph by paragraph. Manual install, ahead of the store.
- VS Code and JetBrains — Masks code and data before they reach the editor AI assistant. Installed from a package, ahead of the marketplaces.
- MCP server — For desktop AI apps and any MCP client, with the same API key.
- Masking API and SDKs — One HTTP call, zero retention. Python and Node libraries on PyPI and npm soon.
- Desktop agent — On-device detection for every application on the machine. Windows, for teams.
Security & privacy
Your data stays encrypted, on infrastructure inside the EU.
filterit adds a privacy and policy layer between your team and AI, helping organisations reduce data exposure and establish responsible AI practices.
Where is the data processed?
Detection processes your text, as you wrote it, on our servers in Germany (Hetzner), together with masking and the vault. The AI provider (Anthropic, outside the EU) receives only the masked text, with placeholders in place of what was detected. Sending anything unmasked takes an explicit choice by you.
Where is it stored?
In our database, inside the EU, encrypted with AES-256-GCM. Real values are not written in readable form. On individual accounts with a password they unlock with your own password (not in organisation workspaces or with Google sign-in). The files you upload are not stored in the database.
For how long?
Files: on temporary disk, deleted within 30 minutes. Placeholder-to-value mappings: 30 days. Conversation history: masked text only, until you delete it or close the account.
What is logged?
Every send to the AI, with destination and time, in a tamper-evident audit trail that holds no personal data. We do not use your data to train models.
- GDPR — Designed for compliance
- Encryption — At rest & in transit
- Egress log — Every send to the AI
- Auditing — Action audit log
Start protecting data today
No credit card · Free to start, Pro and Team when you need them
Frequently asked questions
How do you handle my data?
Detection processes your text on our own infrastructure at a European cloud provider inside the EU (Hetzner, Germany). Sensitive data is stored encrypted and, on individual accounts with a password, under a key that your own password unlocks. We do not use it to train models. The AI model provider (Anthropic) is outside the EU and receives masked text. Every such send is logged and you can inspect it.
Which types of personal data does it detect?
34 categories in total: tax IDs, AMKA, IBANs, names, phone numbers, addresses, ID cards, license plates and more, tuned for Greek data. The full list is on the "Measured accuracy" page.
Does it support Greek documents?
Yes. The engine is built for Greek text, understanding inflections, accents and local data formats, and it is validated on real documents.
Can I integrate it into my own system?
Yes. Through the Masking API with your own key, the MCP server, the VS Code and JetBrains plugins, the browser extension (Chrome, Edge) or the Word add-in. The plugins, the extension and the add-in are installed manually, ahead of the stores. The Python and Node SDKs are published to PyPI and npm soon. See the Developers page.
Is there a free tier?
Yes. You start on the free plan, no credit card, and upgrade only when you need to.