🛡

datenschild

Use AI. Protect data. Analyze company data in a GDPR-compliant way — without personal data ever leaving your organization.

$ echo "User admin@firma.de login von 10.0.0.42" \
  | datenschild --preview

▶ 2 personenbezogene Daten erkannt

User EMAIL_1 login von IP_1

--- Mapping ---
  EMAIL_1 ← admin@firma.de
  IP_1    ← 10.0.0.42

✓ KI sieht nur Platzhalter, nie Echtwerte

Why datenschild?

🔒

GDPR-Compliant

Personal data is automatically replaced with placeholders before AI analysis. The AI never sees real values.

🔄

Automatic De-Anonymization

The AI response is automatically enriched with the original data. You get a readable, complete analysis.

🏠

Local or Cloud

Use Ollama for maximum data privacy (fully local) or Claude API for best quality — you decide.

Zero Dependencies

Pure Python 3.10+ — no external dependencies. Runs instantly, anywhere, without setup overhead.

🔍

10+ PII Categories

Detects emails, IPs, phone numbers, IBANs, addresses, domains, file paths, MAC addresses and more.

📋

Preview & Audit

Preview mode shows you exactly what gets anonymized — before anything is sent to an AI.

How it works

1

Provide input

Pipe log files, emails, or any text via stdin, or load a file with -f.

2

Automatically anonymize

datenschild detects personal data via regex and replaces it with placeholders like EMAIL_1, IP_1.

3

AI analyzes safely

The anonymized text is sent to Ollama (local) or Claude (API). The AI only sees placeholders, never real values.

4

De-anonymize the response

The AI response is automatically merged with the original data. You get the complete result.

Detected data categories

datenschild automatically detects and anonymizes:

EMAIL_n Email addresses
IP_n IPv4 addresses
IPV6_n IPv6 addresses
TELEFON_n Phone numbers
IBAN_n IBAN numbers
ADRESSE_n ZIP + city
DOMAIN_n Domain names
PFAD_n File paths
MAC_n MAC addresses
DATUM_n Dates

Installation

# Clone the repository
$ git clone https://github.com/jriemer/datenschild
$ cd datenschild

# Use directly (no deps required)
$ echo "Log data..." | python3 datenschild.py "Analyse"

# Or install as a package
$ pip install .
$ datenschild -f server.log "Was ist das Problem?"

# Initial setup (choose backend)
$ datenschild --config

Impressum

Angaben gemäß § 5 TMG

Johannes Riemer
Adelbachstraße 14
74405 Gaildorf

Kontakt
E-Mail: jriemer@template7.de

Verantwortlich für den Inhalt nach § 55 Abs. 2 RStV
Johannes Riemer

Close

Privacy Policy

1. Controller
Johannes Riemer, Adelbachstraße 14, 74405 Gaildorf
Email: jriemer@template7.de

2. Hosting
This website is hosted by netcup GmbH (Karlsruhe, Germany). The hosting provider automatically collects information transmitted by your browser in server log files (IP address, browser type, timestamp). This data cannot be attributed to specific individuals. Legal basis: Art. 6(1)(f) GDPR.

3. Analytics
This website uses Matomo (self-hosted, no cookies, no data transfer to third parties) for anonymous usage statistics. You can opt out at any time at analytics.wolkenos.de.

4. Your rights
You have the right to access, rectification, erasure, restriction of processing, data portability, and objection. Contact: jriemer@template7.de

Close