Security & compliance
Trust is an architecture.
Your strategic queries, competitive analysis and regulatory intelligence are sovereign assets. This page says what protects them, in four plain categories: implemented, configurable, planned, and not claimed.
Built with security-conscious engineering practices. Formal compliance attestations are not currently claimed.
Data flow
How data moves, and where the boundaries are.
- Visitor or client uses a browser.
- Requests pass Cloudflare Turnstile and rate limits at intake.
- Server-side functions on Vercel run the logic; it is not exposed to the browser.
- Supabase stores data with row-level security; Resend sends email; Stripe handles payments, and card numbers never reach AimwellBio.
Controls
What is implemented, configurable, planned, and not claimed.
| Area | Status | Detail |
|---|---|---|
| Isolation | Implemented | Client workspaces are isolated at the database layer. Tenant data is logically isolated, with row-level security (RLS) enforced. |
| Authentication | Implemented | Supabase-managed authentication with RLS enforced at the database layer. No credentials are stored in application code. Member API requests require authenticated tokens. |
| Public intake | Implemented | Lead capture and FHIN application routes are token-gated with Cloudflare Turnstile and rate-limited. No unauthenticated route has write access to member or client data. |
| Secrets | Implemented | Credentials are encrypted environment variables in Vercel’s secret management, never in source code, HTML or client-accessible files. Deployment checks look for secrets in public assets. |
| Source lineage | Implemented | Every Cortex output carries source attribution, a reasoning path and a confidence score. |
| Encryption | Implemented | Subscriber data is encrypted at rest (AES-256) and in transit (TLS 1.3). No subscriber data is sold or shared with third parties. |
| Access control | Implemented | Role-based access control across tier levels; administrative access requires multi-factor authentication; least privilege applied. |
| Audit logging | Implemented | Authentication events, API access and administrative actions are logged, with logs retained for at least 90 days. |
| Log export | Configurable | Audit logs are exportable on enterprise request. |
| Data residency | Configurable | Residency configurations are available for enterprise and sovereign deployments, agreed in onboarding. |
| BAA | Configurable | A Business Associate Agreement is available on request for enterprise tiers. |
| SOC 2 Type II | Planned | In pursuit. Current controls include access management, encryption, audit logging, incident response and change management. Audit timing is confirmed on request. |
| Penetration test | Planned | An independent third-party test is planned. A results summary is to be provided to enterprise subscribers under NDA. |
| Certifications | Not claimed | AimwellBio is not SOC 2 (Type I or II) certified, not HIPAA-certified or HITRUST-certified, and not ISO 27001 certified. |
| PHI | Not claimed | AimwellBio does not store or process Protected Health Information in the current platform version. |
| Compliance guarantees | Not claimed | AimwellBio does not provide legal compliance guarantees. Each organization is responsible for its own compliance assessment. |
Infrastructure vendors hold their own attestations (Vercel and Supabase: SOC 2 Type II; Stripe: PCI DSS Level 1). Those are the vendors’ certifications, not AimwellBio’s.
Questions
Answers to the questions investors ask.
Is your source code publicly visible?
No. AimwellBio’s backend infrastructure runs on Vercel serverless functions: compiled, deployed, and inaccessible to the public. No source maps are exposed. No API routes return internal logic. All proprietary intelligence algorithms run server-side. What clients and visitors receive is rendered output only.
What happens if a third-party automation tool fails?
Nothing critical breaks. AimwellBio’s architecture writes to its own infrastructure (Supabase, Resend and Stripe) first. Third-party automation tools are optional mirrors, not primary systems. Platform reliability does not depend on any single external vendor.
How are API keys and credentials protected?
All credentials are stored as encrypted environment variables in Vercel’s secret management system, never in source code, HTML, or client-accessible files. Security audits are run on every deployment to verify no secrets appear in any public-facing asset. Webhook URLs are server-side only.
Testing and disclosure
Report a vulnerability.
Now: dependency vulnerability scanning on every deployment, no secrets in source or public assets (enforced by deployment checks), and OWASP Top 10 coverage through automated static analysis.
Responsible disclosure: if you discover a potential vulnerability, report it to security@aimwellbio.com before any public disclosure. We commit to acknowledging reports within 48 hours and providing a resolution timeline within 7 business days. No legal action against good-faith reporters.
Security testing scope and any penetration-test summary are made available to qualified enterprise and institutional subscribers under NDA on request. Enterprise security reviews are scheduled by request.
AimwellBio provides analytical and informational outputs to support organizational decision-making. It does not constitute medical advice, regulatory guidance, investment recommendations, or legal counsel. All intelligence outputs should be reviewed by qualified professionals before action is taken. AimwellBio is not a registered investment advisor, medical device, or regulatory authority.