Built for real protection
Redbax works at the OS level to intercept, classify and control data before it reaches an AI service — across browsers, desktop apps and IDEs. Everything is inspected locally on the endpoint. No cloud proxy. No browser extensions.
Endpoint Detection Engine
Runs natively on Windows and macOS. Intercepts prompts and file uploads at the OS level before they reach AI services.
- PII (names, emails, phone numbers, national IDs)
- PCI (credit cards, bank accounts, sort codes)
- Credentials (API keys, tokens, connection strings)
- Confidential IP (source code, internal docs, trade secrets)
Powered by a fine-tuned NER model with pattern matching fallbacks. Runs locally — your data never leaves the device for classification.
Policy & Rules Engine
Granular control over what gets blocked, warned, or allowed. Configure policies by data category, team, or service.
- Per-category actions (block PII, warn on source code, allow public data)
- Exclusion lists for trusted domains
- Package security policies — block imports with known CVEs
- AI service allow/deny lists
SOC Dashboard
Full visibility into what your team is sharing with AI tools. Real-time monitoring with actionable insights.
- Live event timeline with category, severity, and action taken
- Per-device and per-user activity views
- Trend analysis and detection volume charts
- Export and API access for SIEM integration
Enterprise-Ready
Built for organisations that take data security seriously. Multi-tenant architecture with full regulatory compliance.
- Multi-tenant isolation: dedicated databases per org
- GDPR, UK DPA 2018, CCPA/CPRA compliance
- Configurable data residency per tenant
- Zero-trust auth with OPA-enforced permissions
- No raw content stored — metadata only