Vatel AI
Company · Trust center

Enterprise-grade voice AI built on trust, security, and sovereignty.

Review Vatel AI's active security posture, SOC 2 alignment, Quebec Law 25 compliance, and data protection controls.

Executive summary

At Vatel AI, privacy and security are built into how we design and operate our voice platform. Call audio, transcripts, and database records are encrypted in transit (TLS) and at rest (AES-256). Our platform runs on SOC 2 Type II certified infrastructure providers (Google Cloud and Supabase), and customers can configure retention rules from 24 hours to 2 years, or disable call recordings entirely. Vatel's own SOC 2 Type II and ISO 27001 programs are currently in progress.

Regulatory standards

Compliance frameworks

Quebec's Law 25

Our practices are designed around Law 25: consent workflows, customer data isolation, configurable retention and deletion, and a designated person in charge of the protection of personal information.

SOC 2 Type II

In progress. Our infrastructure is hosted on SOC 2 Type II certified providers (Google Cloud and Supabase).

PIPEDA (Canada)

Canadian federal privacy standard.

ISO 27001:2022

In progress, aligned with ISO/IEC 27001:2022.

HIPAA

Safeguards for protected health information when healthcare customers use Vatel.

GDPR

Access and erasure rights supported. DPA available at vatel.ai/legal/dpa.

Platform architecture

Core security controls

Data in transit

TLS

Encryption across all SIP trunks, WebSockets & REST APIs.

Data at rest

AES-256

Databases, audio files, and transcripts encrypted at rest.

Access & identity

RBAC

MFA, SSO, need-to-know staff privileges.

Isolated enterprise cloud nodes

Google Cloud & Supabase · Canada / US

Automated retention & destruction · Law 25 / HIPAA

In transit

All telephony SIP traffic, WebRTC audio streams, REST API queries, and WebSockets are encrypted using TLS.

At rest

Database tables, conversation logs, customer records, and call audio files are encrypted using AES-256 bit encryption.

No training on your data

Vatel does not use identifiable customer data to train AI models, and our default language model provider does not train on your data.

Zero data retention (on request)

For customers with stricter requirements, such as healthcare, legal, and financial services, Vatel offers a Zero Data Retention configuration. When enabled, call audio, transcripts, and generated speech are not retained by our speech and voice providers after processing. Contact security@vatel.ai to enable it.

Custom retention schedules

Enterprise administrators can configure custom data retention rules per agent—from 24 hours up to 2 years—after which call recordings, audio snippets, and transcripts are permanently purged from server memory.

Sensitive call handling

Audio recordings can be completely disabled for specific sensitive agents or call types (e.g., medical triage, payment confirmation) while retaining only structured metadata.

Role-based access control (RBAC)

Granular permission controls restricting access to agent configuration, transcripts, and analytics based on employee roles.

Mandatory multi-factor authentication (MFA)

Enforced across all internal administrative access and customer console accounts.

Secrets management

API keys, database credentials, and SIP tokens are managed in isolated, encrypted vaults with automated rotation.

Environment segregation

Strict separation between Development, Staging, and Production environments. Production data is never used in non-production testing.

DDoS & cloud firewalls

Protected by enterprise Web Application Firewalls (WAF) and Cloudflare rate-limiting.

Same-day rollback

Automated CI/CD pipeline allowing instant version rollback in the event of software anomalies.

For procurement

Documents

SOC 2 report

Infrastructure audit

Security questionnaire

Vendor assessment

Third parties

Subprocessors directory

Vendors that may process customer data to operate Vatel. Locations are the primary processing regions for each provider.

CompanyPurposeLocation
Infrastructure
Google CloudServers.United States, Canada
VercelHosting. LLM.United States, Canada
SupabaseAuthentication, database, storage.United States
GitHubSource control, CI/CD.United States
Better StackLogging.United States
StripeBilling.United States
ResendEmail.United States
Voice & audio
ElevenLabsVoices.United States, EU
DeepgramTranscriber.United States, EU
OpenAILLM, Voice.United States, EU
AssemblyAITranscriber.United States, EU
GoogleLLM, Transcriber, Voices.United States, EU
GladiaTranscriber.United States, EU
SonioxTranscriber.United States, EU
SpeechmaticsTranscriber.United States, EU
xAILLM, Voice.United States, EU
CartesiaVoices.United States, EU
Hume AIVoice.United States
MiniMaxVoice, LLM.China
Fish AudioVoices.China
Inworld AIVoice.United States
LLMs & gateways
GoogleLLM.United States, EU
OpenRouterLLM.United States
AnthropicLLM.United States, EU
Alibaba / QwenLLM.China
Moonshot AILLM.China
Zhipu / Z.AILLM.China
IBMLLM.United States
DeepSeekLLM.China
Communications
TwilioTelephony.United States, EU
Customer integrations
Customer HTTP webhooksCustomer-configured endpoints.Customer-determined

SIP peers

SIP trunks and SBCs peered to Vatel are not Vatel-owned subprocessors unless separately contracted.

Trust & security

FAQ

Answers on Law 25, HIPAA, model training, and how to request SOC 2 reports or a DPA.

Vatel is designed around Law 25's privacy-by-design principles. Customer data, voice streams, and transcripts are encrypted in transit and at rest, retention can be set per agent, and call recordings can be disabled. Our person in charge of the protection of personal information can be reached at privacy@vatel.ai.

Yes. For healthcare providers, dental networks, and medical clinics, Vatel implements appropriate safeguards for protected health information. Contact security@vatel.ai to discuss your requirements.

No. Vatel does not use identifiable customer data to train AI models. We may use irreversibly anonymized, aggregated data to improve the service, and you can opt out at any time by contacting privacy@vatel.ai.

To request formal compliance documentation, execute a Data Processing Addendum (DPA), or submit a vendor security assessment, please contact our security team at security@vatel.ai.

Request documents

Tell us what you need and where to send it.

Documents

Request a security review

Headquarters: Devpro Digital / Vatel AI, Montreal, QC, Canada