Enterprise-grade security

Security you can audit.
Compliance you can demonstrate.

Zelvar is built on enterprise-grade infrastructure with full transparency into how your data is handled.

Security Overview

Encryption

All data encrypted at rest (AES-256) and in transit (TLS 1.3). Supabase infrastructure with managed key rotation.

Access Control

Role-based access control (RBAC). 7 role tiers. Row-level security on every database table. JWT authentication on all API routes.

Data Residency

Data stored in US-EAST-1 (AWS). No data sold to third parties. Candidates own their data.

SOC 2 (In Progress)

We are pursuing SOC 2 Type II certification. Expected completion: Q4 2026. Security whitepaper available on request.

Compliance

FCRA

Full FCRA compliance with 3-layer rejection safeguard, adverse action workflow (WF20), and candidate disclosure rights.

TCPA

AI voice agent identifies itself as AI before every call. Explicit written consent obtained before any automated outreach.

EEOC

Bias audit dashboard with Adverse Impact Ratio (AIR) tracking per the EEOC four-fifths rule.

CCPA / GDPR

Candidate data deletion on request. Data processing agreement available for all clients.

Florida TCPA

AI calling scripts include required disclosure language. All-party consent captured before screening calls.

Sub-processors

View full list →
VendorPurposeData ProcessedRegion
SupabaseDatabase & AuthAll candidate + recruiter dataUS-East-1 (AWS)
Anthropic (Claude)AI screening analysisTranscript text onlyUS
Retell AIVoice calling infrastructureAudio recordings (90-day retention)US
TwilioSMS deliveryPhone numbers, message contentUS
SendGridEmail deliveryEmail addresses, contentUS
n8nWorkflow automationTrigger events only, no PII storedEU/US
NetlifyApp hostingNo personal dataUS
CloudflareDNS & domainNo personal dataGlobal

Responsible AI

Zelvar uses AI for candidate screening. We maintain the following principles:

Candidates are always informed they are speaking with an AI

Humans make all final hiring decisions — AI provides scores and summaries only

AI scores are explainable — every score includes reasoning

Bias audit runs automatically — Adverse Impact Ratio tracked per demographic group

Candidates can request human review of any AI decision

Security inquiries

For security inquiries, vulnerability disclosures, or to request our security whitepaper:

[email protected]