AI Compliance for
Engineering Teams
From local CLI mapping to automated CI/CD enforcement. Deterministic compliance snapshots for the EU AI Act.
Technical Preview
Local signal mapping for developers.
Free Local Usage
- Unlimited Local Engine Scans
- Autonomous AI Discovery (`sentinel discover`)
- JSON / SARIF Technical Mapping
- Article-specific Code Pattern ID
- Community Documentation
Evidence Management
Technical Governance & Pipeline Integrity
Flat monthly team plan
- Everything in Technical Preview
- Audit Integrity Ledger (D1 Record)
- CI/CD Watchtower: Automated Gates
- Versioned Compliance Snapshots
- Persistent Governance Logs
- Priority Technical Support
Sovereign Framework
Multi-repo visibility & readiness.
Procurement & Implementation
- Everything in Management
- Annex IV Document Synthesis (DRAFT)
- Enterprise Scout Inventory Crawler
- Sectoral Technical Policy Packs
- Air-gap / On-prem Execution
- Custom Architecture Review
Product Teams
Evaluate your AI system against EU AI Act requirements locally before shipping.
High-Growth Startups
Enforce compliance at the PR level and generate artifacts for audit readiness.
Regulated Organizations
Standardize policy enforcement and procurement vetting across the organization.
Implementation
How it works
01
Install CLI
Run `npx @radu_api/sentinel-scan` to map your AI manifest to EU AI Act articles locally.
02
Enforce in CI
Integrate the GitHub Action to block non-compliant changes at the PR level.
03
Generate Evidence
Export deterministic SARIF/JSON artifacts for internal review and audit preparation.
Data Residency by Design
Local Execution
Scans run in your local environment or CI runner. We don't ingest your training data or model weights.
Deterministic Artifacts
Every scan produces verifiable evidence artifacts (SARIF/JSON) anchored to your development versioning.
CI-Native Enforcement
Works as a standard step in your existing pipelines. Treat compliance as another passing test.
Regulation-Based Mapping
Consistent compliance mapping derived directly from Regulation 2024/1689 (EU AI Act).
Frequently Asked Questions
For technical implementation details, visit our AI Compliance FAQ.
Is Sentinel a managed SaaS?
Sentinel is developer-first infrastructure. You run the Engine locally or in your CI/CD environment. We provide the technical evidence management layer through the Hub, but your source code remains in your perimeter.
Do I need to upload my model?
No. Sentinel uses zero-egress scans. Only technical markers and metadata required for mapping are processed. Your training data and model weights never exit your internal network.
What does Evidence Management unlock?
The Evidence Management plan enables versioned audit history: the Audit Integrity Ledger, automated PR gating via CI/CD Watchtower, and persistent technical governance snapshots.
When should we choose the Sovereign Framework?
The Sovereign Framework is designed for organizations requiring automated Annex IV documentation synthesis, portfolio-wide AI inventory tracking via Enterprise Scout, and air-gapped execution modes.