Immerse — Last updated August 10, 2026 · Effective Date: August 10, 2026
To deliver Immerse, we work with a small set of trusted third-party vendors — we call them processors — who handle specific pieces of infrastructure, AI, analytics, and payments on our behalf. This page explains who they are and what we use each one for, in plain language.
As a general rule, we minimize what any processor sees: AI requests are routed through our own servers wherever possible so providers never see your name or account details, real-time video is processed live and never stored, and while audio is processed live, some providers may retain it briefly for safety and abuse monitoring (as noted in the table below); no processor trains its models on our data.
| Processor | What we use it for | Data involved |
|---|---|---|
| Amazon Web Services (AWS) | Hosts and runs the Immerse platform — our servers, databases, and file storage, plus the content-delivery network that gets media to you quickly wherever you are. | All application data (primary infrastructure) |
| OpenAI | Generates lessons, evaluates learner responses, and produces text-to-speech audio. All requests are routed through our own servers, so OpenAI never receives your name, email, or account ID — and it does not train on our data. | Interaction content only (no account identifiers); retained up to 30 days for abuse monitoring (or zero retention where agreed); not used for training |
| Google (Gemini Live) | Powers real-time, voice- and vision-based coaching during live practice sessions, including on smart glasses. | Live audio/video streams (processed in real time; may be retained briefly for safety/abuse monitoring; not used for training) |
| Deepgram | Converts speech to text in real time on the mobile app, supporting live lessons and voice-based features, including on smart glasses. | Live audio (processed in real time; not retained by default; not used for training) |
| Microsoft Azure Speech | Assesses pronunciation accuracy within our immersive (VR/desktop) experience. | Speech audio (not retained by default) |
| Snowflake | Our analytics warehouse. Usage and progress data is aggregated here to power reporting and product insights. | Aggregated metrics and usage |
| Mixpanel | Tracks product usage events so we can understand how features are used and keep improving the experience. | Usage events |
| Stripe | Processes payments for direct-to-consumer subscriptions purchased through our web or mobile app. | Payment data (held by Stripe) |
| Meta (Horizon Store + Wearables) | Processes payments for subscriptions purchased through our immersive VR experience, and provides the Wearables Device Access Toolkit that lets our app use the camera and microphone on Meta AI glasses for real-time coaching. | Payment data (held by Meta); on-device camera/microphone access for glasses |
| Sentry | Monitors our mobile app for crashes and technical errors so our team can spot and fix issues quickly. | Crash reports and device/session metadata |
| Langfuse | Helps us monitor and continuously improve the quality of our AI-generated content by tracing prompts and responses. Processed in the European Union (Germany). | Prompt/response traces |
This list reflects our current vendors and will be updated as our tools evolve. If you have questions about how any of these processors handle data, reach out to your Immerse account team.