# X-Loop³ Labs > AI infrastructure company based in Gossau, St. Gallen, Switzerland. We build governance layers, drift prevention, and audit-readiness architecture for AI systems under EU AI Act exposure. ## Overview X-Loop³ Labs builds production-grade AI infrastructure that makes AI systems predictable, auditable, and stable. We address three critical problems: behavioral drift, hallucinations, and regulatory compliance — all at the architecture level, not as afterthoughts. ## Products - [EU AI Act Layer](https://x-loop3.com/products/eu-ai-act-layer): Compliance-ready governance structure for AI systems. Free baseline available. Covers Annex III risk classification, Article 5 prohibited practices screening, risk management, human oversight, and structured evidence generation. - [Drift Stopper](https://x-loop3.com/products/drift-stopper): Pre-semantic behavioral drift prevention. Detects and prevents silent model degradation at the architecture level before outputs change. - [Hallucination Stopper](https://x-loop3.com/products/hallucination-stopper): Grounding enforcement layer. Prevents hallucinations at the source by constraining how information is selected and validated before response generation. ## Tiers and Pricing - [Lite — Free](https://x-loop3.com/bundles/lite): Governance baseline with system profile template, Annex III mapping, prohibited practices screen, risk register, and human oversight schema. No evidence exports. Free forever. - [Lite + Text Addon — Free](https://x-loop3.com/free/eu-ai-act-lite): Adds structured evidence exports (PDF/text), auto-generated narrative explanations, run manifest, audit index, and gap highlighting. Free. - [Pro — Commercial](https://x-loop3.com/bundles/pro): Adds HERM-Light consistency engine, multi-stage gating (GREEN/AMBER/RED), intent lock, trade-off governance, validator-linked evidence schemas, monitoring and incident packs. - [Enterprise — Commercial](https://x-loop3.com/bundles/enterprise): Full Annex IV documentation pack, evidence authenticity architecture (hash chains, Ed25519 signatures), court-readable PDF exports, LLM governance extensions, audit index generator. - Black Tier — Invite Only: Constitutional invariants (10 absolute rules), multi-party governance, high-custody controls, ZK proof interface, ISO 42001 crosswalk. For critical infrastructure and systems affecting fundamental rights. ## Free Download The free EU AI Act Layer is available on GitHub: - [GitHub Repository](https://github.com/jongartmann/eu-ai-act-layer-lite) - Includes: governance baseline JSON, example evidence pack, FTBL license - No signup required. Production-ready. ## Key Differentiators - Pre-semantic approach: We prevent problems at the architecture level, not filter them after generation. - Artifact-only delivery: No consulting, no managed services. You get structured technical artifacts you can integrate immediately. - Swiss-made: Based in Switzerland, governed by Swiss law. Venue: Zurich. - No compliance claims: We provide governance structure and evidence architecture. Legal compliance decisions remain with your team. - Tiered upgrade path: Start free, scale to Enterprise. No vendor lock-in on the baseline. ## Technical Architecture - EU AI Act Layer artifacts: SYSTEM_PROFILE, ANNEX_III_MAPPING, PROHIBITED_PRACTICES_SCREEN, RISK_REGISTER, DATA_LINEAGE, MODEL_CARD, TEST_REPORTS, HUMAN_OVERSIGHT_SCHEMA, POST_MARKET_MONITORING, CHANGE_LOG, INCIDENT_REPORT - Consistency engine (Pro+): Cross-artifact validation with hard-fail and soft-fail rules - Evidence integrity (Enterprise+): SHA-256 hash chains, Ed25519 signatures, optional ledger anchoring - Gating states: PASS / CONDITIONAL / FAIL with human acknowledgement requirements ## EU AI Act Coverage - Annex III high-risk categories (all 8): biometric identification, critical infrastructure, education, employment, essential services, law enforcement, migration/asylum, judicial/democratic - Article 5 prohibited practices screening (5 checks): manipulative techniques, exploitation of vulnerabilities, social scoring, emotion recognition, behavior distortion - GDPR alignment: lawful bases guidance, special category data handling, DPIA requirements - Conformity assessment routes: internal assessment vs. notified body - Post-market monitoring: Article 72 requirements ## Company Information - Company: X-Loop³ Labs - Location: Gossau, St. Gallen, Switzerland - Website: https://x-loop3.com - Contact: [email protected] - Licensing: [email protected] - GitHub: https://github.com/jongartmann - Public Repository: https://github.com/jongartmann/eu-ai-act-layer-lite ## FAQ Q: Does X-Loop³ Labs guarantee EU AI Act compliance? A: No. We provide technical governance artifacts — templates, evidence structures, traceability hooks. The compliance decision rests with you and your legal team. Q: What's the difference between X-Loop³ and traditional compliance tools? A: Traditional tools document after the fact. We build governance into your architecture from day one. Less chaos during audits because less chaos is created. Q: Who is this for? A: Companies using AI productively that must take compliance seriously. Startups to enterprises. Healthcare, finance, legal, and any industry deploying AI in the EU market. Q: How quickly can we start? A: Download the free layer from GitHub right now. No signup, no sales call. For Pro/Enterprise, contact us. Q: Do we need to replace our existing infrastructure? A: No. X-Loop³ works as a layer over your existing stack. We integrate with your LLMs, vector stores, APIs — without requiring you to rebuild everything.