Built on trust,verified by design.
Every workflow Zyverno automates runs through the same standard data handled with care, decisions you can explain, and a platform your security team will actually approve of.
This site incorporates information from O*NET Web Services by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). O*NET® is a trademark of USDOL/ETA.
COMPLIANCE
The legal and operational ground rules.
How Zyverno protects candidate and recruiter data, the regulations we follow, and who to contact when you need answers.
Data protection
EnforcedCandidate and client data is encrypted at rest and in transit, stored in access-controlled environments, and never used by Zyverno to train or fine-tune AI models. You decide what's retained and for how long.
GDPR compliance
CertifiedZyverno is built around GDPR principles by default: lawful basis for processing, the right to access or erase candidate data, and clear consent flows baked into every screening and chat interaction.
Security policy
ActiveRole-based access, single sign‑on, audit logs on every record change, and routine penetration testing keep your hiring data isolated and traceable, end to end.
IT support
24/7A dedicated technical team is available for onboarding, integrations, and incident response, so security questions get answered by people, not tickets that disappear into queue.
Transparency
By defaultEvery AI recommendation comes with a visible reason behind it. No hidden scoring, no black-box rejections. Recruiters and candidates can both see what shaped a decision.
DATA CONFIDENTIALITY REQUESTS
Questions about data handling, retention, or your rights as a data subject reach us directly.
dsar@zyverno.app
SECURITY & IT
Reporting a vulnerability or need a technical security review for procurement?
security@zyverno.app
SECURITY
What's actually
protecting your pipeline.
Security pages are usually boring. We tried to keep this one clear: what we protect, how our AI thinks, and the questions we hear most often.
Infrastructure
Hosted on SOC 2 Type II and ISO/IEC 27001 compliant infrastructure, with automated backups and a 99.9% uptime target.
Access control
SSO and role-based permissions mean the right people see only the candidates they need to see.
Candidate data rights
Candidate data can be exported or erased on request, and deleted candidates are anonymised automatically after the retention period.
Reliability
Regular penetration tests, dependency scanning, a published disaster recovery and business continuity plan, and an incident response process keep us proactive.
AI ETHICS
Our AI is an assistant, not a judge.
AI supports decisions. It does not make them.
Candidates are evaluated on skills, experience and role fit. Never on protected characteristics.
Every score is explainable. Recruiters can see why a profile was ranked the way it was.
We test Lina's CV scoring and text screening for bias every year and before we change a scoring prompt or its main model, using candidates who differ only by name.
Infrastructure compliance standards
Looking for enterprise solutions?
Discover how manufacturers automate operations with Zyverno.
FAQS
Common questions about security
Zyverno runs on SOC 2 Type II and ISO/IEC 27001 compliant cloud infrastructure. Files such as CVs are stored in the EU, in Paris. Contact us for the full list of sub-processors and where each one hosts data.
No. Zyverno surfaces ranked candidates, insights and compatibility scores. A human recruiter or hiring manager always reviews and decides who moves forward. This is our 'human-in-the-loop' principle.
Your team can trigger a deletion request from the candidate profile. Zyverno removes the record from active databases and schedules it for purge from backups within 30 days, in line with GDPR retention limits.
Yes. GDPR compliance tools are included across all plans. Our Business plan includes advanced audit tooling and additional data protection controls for regulated industries.
Zyverno does not build its own AI models: Lina's text features use pre-trained models from Anthropic, Google and OpenAI. Before Lina scores a CV or an interview, the candidate's name, email, phone, address and date of birth are removed from the text the scoring model reads. A documented bias test (identical candidates that differ only by name, in several countries) runs every year and before we change a scoring prompt or its main model.
No. Zyverno does not train or fine-tune AI models on your data or your candidates' data. Language model requests go through the Vercel AI Gateway to Amazon Bedrock, Google Vertex AI and OpenAI, with inference in the EU. We require zero data retention on every request, so these providers do not store your request or the answer, apart from limited exceptions in their terms, such as abuse monitoring and short-lived caching. After review by our team, lessons drawn from voice interviews and Lina conversations can improve Lina's written instructions for all customers; the models themselves never change. Voice dictation, voice interviews and search embeddings are handled differently: their providers are listed in our sub-processor register, and the guarantees above do not fully apply to them.
Ready to make hiring effortless?
See how Zyverno unifies your entire recruitment workflow into one calm, intelligent platform, in a 30-minute personalized walkthrough.