Trust Center
Facts, not claims.
What Vellaci does to protect undisclosed vulnerabilities, incident records, SBOMs and security evidence. Everything on this page describes controls that exist today; certifications appear here only once obtained.
Security
Tenant isolation
Row-level security on every tenant table; permissions evaluated in Postgres and re-checked server-side before every mutation; automated cross-tenant tests run in CI.
Authentication
Email/password (≥ 12 chars), TOTP MFA with single-use recovery codes, organisation-level MFA enforcement at page and action level, SAML 2.0 / OIDC SSO with verified domains, JIT provisioning and role mapping, password login can be disabled per domain.
Authorisation
Seven roles with a seeded permission matrix, time-bound support access granted by the customer, time-limited read-only auditor access, partner access with customer-controlled maximum role, two-person approvals for high-risk operations.
Encryption
TLS 1.2+ in transit; AES-256 at rest (managed by the database/storage provider); integration secrets, webhook secrets and SSO client secrets encrypted with AES-256-GCM under a server-held key; API tokens stored as SHA-256 hashes only.
Audit logging
Append-only, hash-chained audit log per organisation with actor, entity, before/after and request context; exports carry chain verification results; CRA submissions are frozen as signed snapshots with amendment history.
Application security
Strict CSP, security headers, schema validation, SSRF guards on every customer-supplied URL, rate limiting on authentication, uploads, search and the API, signed inbound webhooks with replay protection, resource caps on SBOM parsing, no shell or code execution.
Data minimisation
Chat/pager notifications carry identifiers only unless an admin opts in; analytics carry identifiers and counts; the optional AI assistant receives allow-listed, redacted fact sheets — never attachments, source code or secrets.
Due diligence
What you can verify before the first call
The checks an enterprise security, procurement or compliance reviewer usually runs first, and where each answer lives. Anything not public is shared under NDA.
| Question | Where to check | Status |
|---|---|---|
| Who is the provider? | Imprint | Legal entity details not yet published on the site; available on request from hello@vellaci.ch |
| How is tenant data isolated and protected? | Security architecture | Public summary; RLS policies and test evidence under NDA |
| Where is data hosted, and by whom? | Hosting and subprocessors | Public |
| How do we report a vulnerability in Vellaci? | Disclosure policy and security.txt | Public |
| Is the service up, and what happened before? | Status and incident history | Public, with the probe counts behind each figure |
| What changed recently? | Changelog | Public, security-relevant changes marked |
| Can we integrate and export? | API docs and OpenAPI document | Public |
| How is the regulatory content produced? | Content methodology | Public |
| Certifications? | Certifications and readiness | None claimed; programme aligned with SOC 2 / ISO 27001 practices |
| DPA, questionnaire, pen-test summary? | Request under NDA | On request |
Infrastructure
EU-first hosting
EU-first hosting. Database, authentication and object storage run in an EU region; the application tier runs on an EU-region serverless platform.
| Component | Region | Provider |
|---|---|---|
| Postgres database | EU (Frankfurt, eu-central-1) | Supabase |
| Object storage (SBOMs, evidence, reports, exports) | EU (Frankfurt, eu-central-1) | Supabase Storage |
| Authentication | EU (Frankfurt, eu-central-1) | Supabase Auth |
| Application / API | EU (fra1) with EU-only function execution | Vercel |
| Product analytics | EU (eu.i.posthog.com) | PostHog |
| Transactional email | EU sending region where available; content contains no vulnerability details | Resend |
| Billing | Global (card data never touches Vellaci) | Stripe |
| AI assistant (optional, opt-in per organisation) | Anthropic API; minimised fact sheets only; can be disabled entirely | Anthropic |
Dedicated single-tenant hosting (separate project, keys and region) is available on Enterprise contracts. We do not claim sovereign hosting.
Privacy
Data handling
Controller / processor
Vellaci processes customer data as a processor under a data processing agreement; product data (vulnerabilities, incidents, evidence) belongs to the customer.
Personal data
Limited to account data (name, e-mail, role, activity timestamps, IP/user agent in the audit log) and any personal data customers place in records.
Retention
Customer-configurable retention categories; CRA documentation retained ≥ 10 years by default; lapsed subscriptions keep data readable and exportable; organisation deletion has a 30-day cooling-off period with a final export.
Portability
Full organisation export (JSON per table, audit CSV, evidence index, integrity manifest) at any time; public API and CLI.
Compliance
Certifications and readiness
Vellaci operates an information-security programme aligned with SOC 2 and ISO/IEC 27001 practices (access reviews, asset inventory, change management, incident response, backup testing, vendor management, security training, risk register). No certification has been obtained yet; we do not claim one.
Responsible disclosure
Report a vulnerability in Vellaci
E-mail security@vellaci.ch or use the disclosure form. We acknowledge within 2 business days and coordinate disclosure with reporters. Machine-readable contact: /.well-known/security.txt.
Subprocessors
Who processes data on our behalf
| Subprocessor | Purpose | Location |
|---|---|---|
| Supabase | Database, authentication, storage | EU (Frankfurt) |
| Vercel | Application hosting and CDN | EU (Frankfurt) functions; global edge network for static assets |
| Stripe | Subscription billing | EU/US (PCI DSS Level 1) |
| Resend | Transactional email | EU/US |
| PostHog | Product analytics | EU (Frankfurt) |
| Anthropic | AI assistant (only when enabled by the organisation) | US/EU API endpoints; minimised data |
| Atlassian / Linear / Slack / Microsoft / PagerDuty / GitHub / GitLab | Customer-configured integrations | As configured by the customer |
Policies
Documents available under NDA
- Threat model and security controls
- Penetration test scope and latest summary
- Incident response runbook
- Backup and disaster recovery plan (RPO/RTO)
- Business continuity for CRA deadlines
- Security questionnaire pack (CAIQ-style)
- Data processing agreement and subprocessor register
- Access review and change management procedures
Running a security review?
We share architecture details, RLS policies, test evidence and the questionnaire pack under NDA.