Govern your AI the way the EU AI Act demands. In the product, not in a PDF.
Risk-based AI system inventory, versioned FRIA, incidents with the 15-day clock and every model call backed by a verifiable evidence chain. The EU AI Act and ISO 42001, made real.
No credit card required. Free plan to get started.
The problem
AI is already in production. Governance is not.
You don't know what your team sent to OpenAI yesterday
Without a record of every call —who, when, with what data— AI in production is a black box you cannot audit. And the AI Act audits systems, not good intentions.
High-risk with no impact assessment = an audit finding
Since 2 February 2026, deploying high-risk AI in the EU requires a FRIA (Art. 27). If yours lives in a Word file in a shared folder, it does not exist for the auditor.
A log you can edit is not proof
When the regulator asks "prove this was not tampered with," a record in mutable rows is worthless. You need evidence, not a rewritable history.
AI governance, made real
Every capability anchored to a real obligation under the EU AI Act, ISO 42001 and ISO 38507.
Risk-based AI system inventory
Every AI system —internal or external— catalogued with its lifecycle and its EU AI Act risk class (prohibited / high / limited / minimal). The subject everything else hangs from.
Versioned FRIA with separation of duties
The fundamental-rights impact assessment as a versioned file bound to the AI system, with four-eyes approval (author ≠ approver) and an automatic finding when one is missing.
Incidents with the 15-day clock
A serious incident starts the Art. 73 counter and blocks closure until root cause, corrective actions and the notification to the authority are on record. The deadline no longer depends on someone remembering.
PII redaction on the critical path
Before an ID number, IBAN or card reaches OpenAI or Groq, it is redacted automatically —in the same call— and the detected categories are logged. Multi-jurisdiction coverage, by default.
Model & Dataset Cards with status
Intended use, limits and known biases for each model, plus your RAG corpus documented — versioned, with a warning when they go stale.
Audit trail with a verifiable evidence chain
Every AI call is cryptographically chained by a database trigger. We never store the prompt: we store the proof that you can demonstrate it. Verifiable with one click.
Precios
Empieza gratis. Escala cuando lo necesites.
- Hasta 50 assets
- 1 fuente de datos
- Lineage a nivel tabla
- Auto-documentación básica
- 3 usuarios
- Hasta 500 assets
- Conectores ilimitados
- Glosario de negocio
- Quality rules
- Designer visual
- Governance y ownership
- Email support
- 10 usuarios
- Hasta 5,000 assets
- Chat IA avanzado
- Lineage a nivel columna
- Conectores ilimitados
- Quality rules automatizadas
- Designer visual multi-plataforma
- MCP Server (+38 herramientas)
- Priority support + onboarding guiado
- 50 usuarios
- Todo lo de Professional
- Assets y usuarios a medida
- SSO/SAML
- AI providers personalizables
- Lineage + impact analysis avanzado
- Integraciones PM (Jira, Linear, GitHub)
- RBAC multi-nivel
- Soporte dedicado
FAQ
Respuestas sobre implementación y capacidades
Your AI, governed from day one.
The EU AI Act is already law: FRIA, the Art. 73 clock and high-risk logging will not wait. Try it in your own environment and verify it yourself.