Cybersecurity Made in Europe
Labelled under the ECSO scheme, issued by the European DIGITAL SME Alliance.
A declaration-based label, not a certification or audit.
Built to be checked, not believed.
Every answer cites or refuses. Limits and prices are published. Here is how Ansvar handles your data, where it runs, and where our audits stand.
Passed BSI's ISO/IEC 27001:2022 certification audit, November 2026. Certificate to follow. Request the audit report under NDA
Your questions are not stored by the corpora.
Corpora never store client data
Corpus servers answer from law, regulation and standards. Your question is not stored there.
Documents stay tenant-isolated in the EU
Documents you register on Team and Company live in tenant-isolated EU storage under your retention policy, and the same tools that register a document can remove it.
EU core
Hetzner, EU. Edge and backups are disclosed on the subprocessors page. No tracking cookies on this site.
Your model, your provider
Ansvar serves evidence to the AI client you already run and never sees model traffic. Customer data is never used for training. The model and its provider keep their own terms.
You choose where Ansvar runs.
Pick a setup to see who runs it, where your data lives and what crosses your boundary.
Your organisation
Your teamYour systems
Questions and answers, over the internet
Ansvar, EUrun by us
Law and standardsAssessments
Your workspace, kept separate
Your documentsYour reports
Your AI provider processes your prompts and answers under its own terms
- Your documents live
- In your own workspace in our EU cloud, kept apart from every other customer.
- Who runs it
- We do.
- What crosses your boundary
- Your AI's questions and Ansvar's answers.
- Updates
- Continuous, by us.
- Not yet in this setup
- Nothing. Answers, assessments, document upload and regulatory updates all run here.
What changes when it moves.
| Aspect | Our EU cloud | Your own cluster | Air-gapped |
|---|---|---|---|
| Where it runs | EU infrastructure, Hetzner | The Ansvar connector, in your cluster | The Ansvar connector, on your premises, no connection out |
| Who runs it | Ansvar | Your team, with our runbooks. We never operate it remotely. | Your team, with our runbooks. We never operate it remotely. |
| Your documents | Your own workspace in our EU cloud, kept apart from every other customer | In your cluster, read from storage you mount | On your premises |
| Law and standards | Served from the central gateway | Served from the central gateway | Signed content packs, carried in by you |
| Assessments | Available | Run on our EU gateway today, not inside your cluster | Not on site yet |
| Document upload | Available | Not yet: the connector reads storage you mount | Not yet: the connector reads storage you mount |
| Chat on site | Not applicable | Not applicable | Not yet |
| Updates | Continuous | Versioned releases you pull and install | Signed packs, on your schedule |
| Licensed standards | Add-on per seat | Add-on per seat, served centrally | Not on site; licensed standard text never ships in a pack |
| Status | Available now | Pilots open soon | Pilots open soon |
| Price | Published plans | Annual contract | Annual contract |
Where the law comes from.
Each link holds for every served row. A source still under licence review is never fetched, so it never enters the chain.
For NIS2 customers assessing us as a supplier.
| Area | What we have in place |
|---|---|
| Security policy | Information security policy and a management commitment approved by the management body, including NIS2 Article 20 oversight |
| Risk management | Risk methodology, risk register, management reviews, and an internal audit performed by an external auditor |
| Incident handling | Incident response procedure with staged report templates for NIS2 Article 23 |
| Continuity and recovery | Continuity and disaster-recovery policies, tested restores and encrypted backups |
| Supply chain | Supplier assessments that cover secure development, vulnerabilities and product quality, as Article 21(3) asks |
| Development and vulnerabilities | Changes through pull requests, organisation-wide code and dependency scanning, and coordinated vulnerability disclosure |
| Effectiveness | Measured controls, internal audit and management review |
| Hygiene and training | Security awareness training for all staff and the management body; endpoint baseline with attestation |
| Cryptography | Cryptography policy, TLS 1.2 or higher, and encryption at rest for the databases holding customer data |
| People | Screening, security terms in contracts, recorded offboarding |
| Access control | SSO, MFA on privileged accounts, tenant isolation and quarterly access reviews |
| Assets | An asset inventory reconciled against live systems |
| Physical security | No premises of our own; hosting in ISO 27001-certified Hetzner data centres in the EU |
Our DPA covers breach notification and audit rights. If your NIS2 supplier assessment needs more, we offer a security addendum with incident reporting stages aligned to Article 23, threat notices and evidence rights.
Where we stand, in plain words.
ISO/IEC 27001:2022: we passed BSI's certification audit in November 2026. Until the certificate is issued, we share the audit report under NDA on request.
We run an AI management system (AIMS) as part of our ISO 27001 management system. It covers the AI systems we build and operate. We plan ISO/IEC 42001 certification for the second quarter of 2027.
Ansvar is too small to be a NIS2 entity itself today. Many of our customers are, so we work to what their supplier checks under Article 21(2)(d) ask for: ISO 27001 controls, coordinated vulnerability disclosure, audit rights, and a security addendum with incident reporting stages aligned to Article 23. The NIS2 assurance pack maps our controls to Implementing Regulation (EU) 2024/2690.
The Cyber Resilience Act covers products with digital elements. Our hosted service is not one. The package we deliver for your own environment is, and we build it as one: every release ships signed, with a software bill of materials, vulnerabilities are handled under the disclosure policy below, and we will report actively exploited vulnerabilities as Article 14 requires.
A white-box test of the platform by our own team in August 2026. An independent third-party test is planned; we publish its status when a report exists.
Our ISMS is run by a team holding ISO/IEC 27001 Lead Auditor and Lead Implementer certifications.
Labels and public commitments
These describe our declarations and commitments. They are not independent security audits.
CSA STAR Level 1
Listed in the CSA STAR Registry: Ansvar Gateway’s CAIQ v4.0.3 self-assessment.
A published self-assessment, not a third-party certification.
EU AI Pact
Signatory of the European Commission’s AI Pact, Pillar II pledge.
A voluntary pledge, not a certification or audit.
Report a vulnerability.
Our Terms otherwise prohibit probing the Service, so this section is the permission: we authorise good-faith security research on the systems below, and we will not take legal action over research that follows this policy.
Email security@ansvar.eu. PGP key on request.
- ansvar.eu, app.ansvar.eu, and gateway.ansvar.eu
- Reaching another tenant's data, audit logs, organisation membership, or billing, whether you can read it or change it
- Authentication and authorisation flaws: OAuth and MCP token confusion, session leakage, tier or entitlement bypass
- Injection, server-side request forgery, and remote code execution
- Exposed credentials, keys, or customer data, including our own misconfiguration of Cloudflare, Vercel, or Hetzner
- Corpus integrity: source or cache poisoning, provenance tampering, or a bypass of licensing suppression
- Missing security headers, cookie flags, or TLS settings with no demonstrated impact
- SPF, DKIM, and DMARC gaps without impact you can demonstrate on a domain or mailbox you control
- Clickjacking on pages that change no state, and rate limits with no demonstrated security impact
- Version banners, and scanner output with no proof of concept
- Denial of service and load testing, social engineering, physical testing, and flaws in Cloudflare, Vercel, or Hetzner themselves
- Wrong or missing citations in served content. That is a support matter, and rights complaints go through content claims.
- Email security@ansvar.eu. PGP key on request.
- Send what you found, the steps to reproduce it, and the impact as you read it.
- We aim to acknowledge within one business day and commit to three, and we tell you what we decided once we have triaged it.
- Stop at proof. Do not pull customer data, pivot further into our estate, or run load tests.
- If you reach another tenant's data, stop and tell us. That finding is already complete.
- Give us 90 days before you publish, and tell us when you plan to.
- No paid bounty. We would rather say so than imply a payout that is not there.
- A credit in the acknowledgements below, under your name, your handle, or nothing at all. We ask before we publish anything.
- A numbered “I hacked Ansvar” coin, which we make ourselves and do not sell.
- We decide both after triage, at our discretion. Recognition is not compensation, and we do not price a finding before we have read it.
- Stay in scope, use only accounts you own or may use, and avoid disrupting the service.
- Access no more data than proves the finding, keep no copies, and delete what you took.
- Report promptly and agree publication timing with us.
- Do that and we will not start legal action over your research, and we will confirm to others that it was authorised.
- We cannot waive our customers' or licensors' rights, and this binds no public authority.
Acknowledgements
We have not credited anyone yet. Once we have fixed a valid report, we list the researcher here the way they asked to be named.
Questions security teams ask.
No. Ansvar runs no server-side model and never sees model traffic. It flows directly from your AI client to your model provider. Customer data is never used for training, evaluation, or fine-tuning. Your provider's own retention terms apply on that direct path. Where Ansvar operates a run for you, in expert-delivered engagements, the same no-training rule applies, and the model provider's published retention terms govern that path.
Corpus MCP servers store no client data. The gateway holds account records, entitlement records, and operational security logs on EU-hosted Kubernetes infrastructure with NetworkPolicy isolation between services and envelope-encrypted secrets (OpenBao KMS). Document grounding on Team and Company stores registered documents and workflow evidence in tenant-isolated EU storage under your retention policy, and the same tools that register a document can remove it. Company adds a separate per-tenant cryptographic ledger with signed receipts.
We notify affected customer admins without undue delay, per our DPA, early enough to support your own clocks, including your GDPR Art. 33 notification to your authority (72 hours, controller to authority) and a NIS2 24-hour early warning where it applies to you.
Not directly today. Ansvar is below the size threshold, and none of the size-independent categories applies to it. Many of our customers are essential or important entities, so we run our ISMS against the NIS2 Article 21 measures and map it to Commission Implementing Regulation (EU) 2024/2690. The map is on this page; the full assurance pack is available on request under NDA.
SSO: yes. Sign in with Microsoft Entra ID, Google, or username and password. SCIM 2.0 is planned; it is not built today.
Customer-managed deployment is at the design-partner stage. A deployable platform build exists and runs in CI on each merge. Ansvar has not released it as a self-serve product. Ansvar and the first production partner must agree on corpus updates, freshness reporting, support, and licensed-content entitlements. The gateway runtime and base images remain private; Ansvar publishes a small set of connector repositories where licensing allows.
We performed a white-box penetration test of our platform in August 2026, covering the MCP gateway, authentication, and web applications. An independent third-party test is planned; we publish its status when a report exists.
Email security@ansvar.eu, with a PGP key on request. We aim to acknowledge within one business day and commit to three. Our Terms otherwise prohibit probing the Service, so the responsible disclosure section on this page is the permission: it authorises good-faith research, sets the scope, and states the safe harbour that applies if you stay inside it. We run no paid bounty. A valid finding may earn a credit in our acknowledgements and a numbered “I hacked Ansvar” coin, decided after triage.