Try it free in the sandbox · first FHIR call in 5 minutesFree · 30 synthetic patients · no cardYour first API call in 2 minutes →

Trust at Anpheros Platform: what we have, what we don't, what's next

What protects health data on Anpheros Platform today (EU hosting, DPA, encryption, consent, audit), what is missing (no own SOC 2 or ISO 27001 yet, no external penetration test) and what comes next.

Anpheros is in private beta. This page says plainly what protects the data today, what is missing, and when it comes.

What we have

EU hosting. Platform data is stored and processed in the European Union by default (Google Cloud europe-west4, Netherlands; webhook queue in Frankfurt; audit export in the BigQuery EU location). Sub-processors are listed in our data processing agreement.
GDPR roles and a DPA. The clinic or company is the controller; Anpheros is the processor. For data a patient keeps in their own Anpheros app account, Anpheros is the controller. Production keys are issued only to verified organisations with a signed data processing agreement.
US customers. For US production customers we can set up a US data region and sign a HIPAA Business Associate Agreement; Google Cloud's own BAA covers the underlying services.
Encryption. HTTPS only with HSTS; data encrypted at rest by Google Cloud; documents in an EU bucket with a customer-managed encryption key (CMEK), immutable after finalisation, with a SHA-256 checksum.
The patient controls access. Consent per resource type and duration (30–365 days) on a hosted consent page, revocable at once. Revoking stops new access. Records a clinic has already saved in its own system are kept by that clinic under its own legal duties. Every grant is also a FHIR Consent resource.
Audit the patient can read. Every read and write is logged, including reads with a project's own key and context built for AI; the log is hash-chained and exported append-only.
Provenance on every write: who wrote it (patient, practitioner, device, import, AI), for which organisation, from which system.
Isolation between applications. Pairwise identifiers per project; sandbox and production are separate databases, and a test key cannot reach real patients.
Credentials. API keys stored only as hashes; OAuth 2.1 with mandatory PKCE, short-lived rotating tokens with reuse detection; MFA required for organisations with real data.
Infrastructure certifications we inherit from Google Cloud (for the data centres and cloud services, not for Anpheros itself): ISO/IEC 27001, 27017, 27018 and 27701, SOC 1/2/3 reports, CSA STAR. See Google Cloud's compliance offerings.
Public availability: developers.anpheros.com/status, measured every five minutes from outside the service.

What we don't have (yet)

No SOC 2, ISO 27001 or HITRUST certification of our own. Google's certifications cover the infrastructure, not our company's processes.
No database failover across zones. The service runs across zones; the database runs in one zone with point-in-time recovery and backups stored outside the zone.
No contractual SLA with service credits. Paid plans have a best-effort service level (see the status page: targets, not guarantees, no service credits).
No external penetration test yet; security reviews so far are internal.
Private beta: production opens per verified organisation.

What's next

These are our plans, not contractual commitments; the order and timing may change.

With the first paying production customer: a database with automatic failover (high availability).
When there is revenue: an external penetration test, with a public summary.
When a hospital or enterprise customer requires it: ISO 27001 (first, for the EU) or SOC 2 Type I, then Type II.
Continuing: the status page history and published incidents.

Contact

Security questions and reports: [email protected] (privacy: [email protected]).

Security, privacy and data residency

Ask your AI

Paste this into Claude, ChatGPT, Cursor or Claude Code. Your assistant reads this guide and adapts it to your code.

Read https://developers.anpheros.com/guides/trust.md and help me apply this guide to my own project. Use the Anpheros sandbox (base URL https://platform.anpheros.com) with my API key from the ANPHEROS_KEY environment variable, never written in the code. Give me complete, runnable code, the response I should expect, and how to fix the errors in https://developers.anpheros.com/guides/troubleshooting.md if a call fails.

Open in Claude Open in ChatGPT This guide as Markdown Build the whole integration with an AI coding tool