Connect uses the providers below for specific features. Their locations, retention, and onward transfers depend on the service and account configuration; this page is not a certification.
Connect integrates with cloud, voice, analytics, payment, and email providers. Use of a provider is conditional on the feature you enable; provider terms and retention rules may also apply.
The table below lists providers observed in the current code and deployment documentation. Locations must be confirmed against the current provider contract and tenant configuration.
| Service Provider | Service Category | Location | Purpose |
|---|---|---|---|
| Cartesia | Voice Synthesis | United States | High-quality voice generation and synthesis for translated output |
| Soniox | Speech Recognition | United States | Real-time speech-to-text transcription and recognition |
| Sentry | Error & Bug Tracking | United States | Real-time error monitoring and reporting |
| Firebase / Google Cloud | Authentication, Database & Storage | Provider-dependent | Firebase Authentication, Firestore, Cloud Storage, Cloud Functions, and Google sign-in |
| Stripe, Inc. | Payment Processing | United States | Payment processing and subscription billing management |
| Hetzner | Cloud Infrastructure | Germany / Europe | Cloud server hosting and infrastructure for application deployment |
| Vercel Inc. | Web Hosting & CDN | United States | Website hosting, deployment, and global content delivery |
| Google Analytics | Consent-gated Web Analytics | Provider-dependent | Website measurement after the applicable consent choice |
| Amplitude | Product Analytics | United States | Product usage analytics and user behavior tracking |
| Resend | Transactional Email | United States | Sending transactional emails, account notifications, and verification |
| PyannoteAI (optional) | Speaker Identification | Provider-dependent | Speaker identification and voiceprint enrolment when the feature is enabled |
| Gender Recognition (optional) | Audio Inference | Provider-dependent | Automated gender estimate from a short audio sample when the feature is enabled |
Scope note: AIC/Quail models observed in the API are currently processed on Connect infrastructure; no direct transfer of end-user audio to AIC was established. Hume AI, OpenAI, Anthropic, ElevenLabs, and Loops were not observed in the current production path and are not listed.
We may update this list from time to time as we add or remove service providers. All changes will be reflected on this page with an updated date.
For questions about our subprocessors or data processing practices, use our contact page.
Visit our Privacy Policy or contact our support team for more information on how we handle your data.
