# LexITanon > LexITanon pseudonymises and redacts German legal documents (court decisions, orders, briefs) automatically > and entirely on the customer's own computer or server - no cloud service, no external AI interface. It is > made for law firms, courts, publishers and legal-tech or AI products that must remove personal data before > a document is published, shared or sent to an AI model. Made by be next GmbH, Stuttgart, Germany. > Current release: LexITanon 0.5.0 with AI model 1.3.3. The website and the program are in German. ## What it does - Finds personal data with its own AI recognition model trained for German court decisions, plus fixed rules (dates, IBAN, tax ID, phone numbers, e-mail addresses, case numbers ...). The model only marks words; it does not write or invent text and does not learn from customer documents. - Categories: persons, companies and organisations, addresses and places, dates, the case number of the proceedings, contact data, bank accounts, tax ID, ID numbers, health data. Courts, laws, amounts and cited decisions are kept. - Pseudonymisation with placeholders such as [NAME], [ANSCHRIFT], [DATUM], optionally numbered ([NAME_1], [FIRMA_2]; the same person keeps the same number), or redaction with black bars. - Dates: replace, keep, keep only month and year, or shift all dates by the same number of weeks (intervals stay intact; birth dates are always replaced). - Input: PDF with a text layer, UTF-8 text. Output: a new PDF or text. Scanned PDFs need OCR first; password-protected PDFs and Word files are not supported. - Like every automatic method it can miss data or replace too much; every finding is shown for review. ## Versions and prices (19 % VAT added for Desktop and Server) - LexITanon Desktop - one workstation (Windows 10/11, macOS 13+ with Apple chip, Linux x86-64), operated in a local window or browser: 7 EUR per user and month billed yearly, 20 EUR with a monthly licence. - LexITanon Server - the whole firm (Linux x86-64, Windows Server 2016+ as a service): browser access from all workstations, several documents in parallel, REST interface (OpenAPI) for practice-management software, DMS and AI applications, custom rules for the firm's own file numbers ([CUSTOM_n]): 195 EUR per running instance and month billed yearly, 350 EUR with a monthly licence; any number of workstations. - LexITanon Mobile 0.5.0 - iPhone, iPad (iOS 16.4+) and Android 9+: single documents on the device, 4.89 EUR per month incl. VAT as an in-app subscription, or free for one device per Desktop workstation or Server licence. App Store: https://apps.apple.com/de/app/lexitanon/id6830373662 - Google Play: https://play.google.com/store/apps/details?id=com.benext.lexitanon - Licences are annual or monthly and do not renew automatically. Order: https://lexitanon.com/bestellen/ ## Privacy and network - Documents are processed only in memory on the machine where LexITanon runs; no documents, file names, texts or findings are ever transmitted. - The only internet connection is the licence check (Desktop daily, Server hourly): licence key, program version, a hardware checksum and usage counts (numbers only). Without a connection it keeps working for 14 days (Desktop) or 24 hours (Server); offline licences on request. - Optional "Fehlerhafte Erkennung melden" (report a recognition error): sends one example with invented replacement data, only after the user's explicit consent. - NOCIP Gold label (local AI verified), no. G-2026-0001, for LexITanon Desktop for Linux 0.5.0, valid until 09.11.2027: https://nocip.com/register/NOCIP-G-2026-0001/ ## Integration (Server version) - REST API (JSON over HTTP, OpenAPI description built in): pseudonymise a text immediately (POST /api/v1/redact-text), submit PDF or text files as jobs, download the result as PDF, text or JSON with all findings and offsets. English guide: https://lexitanon.com/assets/docs/LexITanon_Server_API_Guide_EN.pdf - Python client and examples (standard library only): https://gitlab.com/beaexpert/lexitanon - lexitanon-mcp (in the Python client): an MCP server for AI assistants and agents. A file is pseudonymised on the customer's LexITanon server; the pseudonymised result then goes to the AI service the user chooses (often a cloud service). No automatic recognition is complete; results should be reviewed. - Typical use: pseudonymise a document locally before it is sent to an AI service (ChatGPT, Copilot, Claude and others), so the AI only sees placeholders. ## Pages - [Home](https://lexitanon.com/): overview, before/after example - [Funktionen](https://lexitanon.com/funktionen/): what is recognised and kept, result, REST interface, limits - [KI & Berufsrecht](https://lexitanon.com/ki-und-berufsrecht/): AI in law firms - section 203 StGB, sections 43a/43e BRAO, Art. 28 GDPR; what the licence check sends - [Versionen](https://lexitanon.com/versionen/): Mobile, Desktop and Server compared, system requirements, current versions - [Preise](https://lexitanon.com/preise/): prices and how to order - [Demo](https://lexitanon.com/demo/): try it in the browser without sign-up (at most 20 findings per text) - [Handbuch](https://lexitanon.com/handbuch/): user manual (PDF, German) and API guide (PDF, English) - [FAQ](https://lexitanon.com/faq/): frequently asked questions - [English summary](https://lexitanon.com/en/): LexITanon in English - [Kontakt](https://lexitanon.com/kontakt/): enquiries; [Impressum](https://lexitanon.com/impressum/): be next GmbH, Kirchheimer Str. 27, 70619 Stuttgart, Germany