Security & Trust
Your data is yours. We never train on it.
BeneDoc is built so you can put confidential data into the platform and know exactly how it is protected: encrypted, US-hosted, used only to serve you, and yours to export and delete at any time.
Protections at a glance
What you can count on
Your data stays yours
We process your data only to provide the service, on your documented instructions, and for no purpose of our own.
Never used to train models
Your data is never used to train, fine-tune, or improve any general model, and it is never sold or shared for advertising.
Encrypted and US-hosted
Data is encrypted at rest with AES-256 and in transit with TLS 1.2 or higher, on Google Cloud in the United States, with access limited to those who need it.
A short, disclosed subprocessor list
Every third party that touches your data is named below, with advance notice and a right to object before the list changes.
72-hour breach notice
If a security incident affects your data, we notify you within 72 hours of becoming aware, with what we know.
Clean exit, on a clock
Request your data back within 30 days of termination and we return it within 30 days. Everything is deleted within 60 days, with a deletion certificate on request.
How your data is handled
Built secure, not secured later
Purpose-limited processing
Your data is used only to run the service you asked for. We do not repurpose it, sell it, or share it for advertising.
Least-privilege access
Access is identity-based and need-to-know. Credentials never live in our code; they are held in a managed secrets vault.
A tamper-evident trail
The platform keeps immutable, append-only audit logs engineered to 21 CFR Part 11 standards, so every change is traceable.
US data residency
Your data is processed and stored in the United States on Google Cloud, whose infrastructure carries SOC 2, ISO 27001, and PCI DSS.
Backups that age out
Routine backups are encrypted and expire within 30 days. They are never restored to production except to recover from a failure or an incident.
AI you can trust with private data
Automated, but never at your expense
No training on your data
Your content is used to produce your results and nothing else. It never becomes training data for any model.
Human-final by design
Outputs are drafts for your review. You apply the final judgment and remain the decision-maker, every time.
Ready for 2026 privacy rules
We support your obligations around automated decision-making and risk assessments under the CCPA/CPRA and comparable state laws.
Traceable to the source
Outputs carry source traceability and citations, so every result can be checked back to its underlying evidence.
Compliance and standing
Where we stand today
CCPA/CPRA Service Provider
For US customers, we contract as a Service Provider under the CCPA/CPRA and comparable state privacy laws: no sale, no sharing, purpose-limited.
SOC 2-aligned program
Our controls are designed to the SOC 2 Trust Services Criteria across security, availability, confidentiality, processing integrity, and privacy.
Ready for the EU when you need it
If your data reaches the GDPR, UK GDPR, or Swiss FADP, our DPA carries the Standard Contractual Clauses and UK Addendum automatically.
Health data under a BAA
We process protected health information under an executed business associate agreement, which governs that data and takes precedence over our other terms.
We answer your questions, on a deadline
Send a security questionnaire to our Security Contact and we respond within 30 days, once a year and again after any incident affecting your data. We provide our documentation under NDA where appropriate.
Our subprocessors
Every third party that hosts, stores, or otherwise accesses customer data. This is Annex III of our DPA.
| Subprocessor | Location | Processing task |
|---|---|---|
| Google Cloud Platform (Google LLC) | United States (us-east4) | Hosting, storage, compute, data warehouse, secrets management |
| Google LLC (Vertex AI) | United States (US multi-region) | Large-language-model inference and embeddings |
| Google LLC (Gemini API) | United States | Model inference for certain platform features, and failover for Vertex AI |
| Anthropic, PBC | United States | Large-language-model inference |
| OpenAI OpCo, LLC | United States | Large-language-model inference and embeddings |
| Stripe, Inc. | United States | Payment processing and subscription billing |
| Resend, Inc. | United States | Transactional email delivery |
A third party belongs on this list only if it hosts, stores, or otherwise accesses customer data. Layers that transport it without accessing it, such as mesh-VPN networking and outbound egress proxies, carry only encrypted or non-customer traffic and are infrastructure rather than subprocessors. Internal engineering tools like our source-code host and issue tracker are excluded too, because we keep customer data out of them.
Where you designate external sources for us to access on your behalf, those are your own selected sources rather than our subprocessors.
We give at least 30 days' notice before adding or replacing a subprocessor. You can object on reasonable data-protection grounds within that window, and if we cannot resolve it you can terminate the affected part of the service and we refund fees prepaid for the period after termination.
Policies and documentation
The formal instruments behind the summary above.
- Privacy Policy: how we handle personal information.
- Terms and Conditions: the terms for using our website and service.
- Data Processing Agreement, our standard DPA, which attaches to each customer agreement. Available on request.
- Information Security Overview, our technical and organizational security controls. Available on request.
Have a security or data question?
Our team answers due-diligence requests and security questionnaires directly.