Anonymize and classify government documents. Zero data persistence.
AI-powered document anonymization and classification for government agencies. Five protection categories. Structured placeholders or full redaction. Zero-persistence architecture, local deployment, FADP/GDPR compliant.
The Challenge
- Freedom of information requests require anonymization
- Manual redaction is slow and error-prone
- Cloud solutions raise data sovereignty concerns
- Multiple document formats need consistent handling
- Strict compliance requirements under FADP and cantonal law
FADP-compliant by architecture
Shield's zero-persistence architecture means personal data is never stored. Documents are processed entirely in memory. This simplifies compliance. There is no data retention to manage, no database to secure, no logs to audit for personal data.
Sovereign deployment
Shield runs on your agency's infrastructure with no external dependencies. Air-gapped deployment requires no internet connection. Docker containers deploy on existing government IT infrastructure. Data never leaves your premises.
Consistent anonymization at scale
AI-powered entity detection produces consistent results across documents and staff members. Smart preservation rules ensure publicly known officials retain their names while private citizens are protected. Review and approve in seconds, not hours.
Frequently Asked Questions
Stop pasting AI output into Word.
Two products. Sphere: a 40-tool legal AI agent that outputs real tracked changes. Shield: court anonymization on a single 24 GB GPU, zero data stored. Swiss engineered.
Explore Shield