Sub-processors
The companies that process data for KardoVision, what each does, and where.
Draft — not yet reviewed by counsel. This text describes KardoVision as it works today. Anything in [brackets] is a placeholder to be settled before this page is final.
Kardo runs KardoVision on the services below and on nothing else. Video recorded on your KardoHub and cameras never goes to any of them unless a camera’s privacy mode sends pictures or clips to the cloud. The two AI providers receive anything only while the corresponding switch on Cloud & Privacy is on. We will update this page, and email account owners, at least [30] days before adding or replacing a sub-processor, as the Data Processing Addendum says.
| Sub-processor | Purpose | Data it receives | Region | When |
|---|---|---|---|---|
| Amazon Web Services (AWS) | Cloud hosting: the API and web servers, the database and cache, object storage (S3) for clips, exports, screenshots and backups, and the device connection service (IoT) that carries each device’s certificate, settings and commands. | All account data and any customer content that reaches the cloud, at rest and in transit. | us-east-1 (Northern Virginia, United States) | Always. |
| Resend | Sending email: sign-up and reset codes, invitations, alerts, daily summaries, security notices — from noreply@kardovision.com. | Recipient email address and the message (alert titles, site names; amounts only for people with that permission; never a picture). | United States (provider-operated) | Always; per-person and per-account caps apply to alert email. |
| Expo push service | Delivering push notifications to phones running the KardoVision app (door calls, alerts). | The phone’s push token and the notification text (a door name, an alert title, a site). | United States (provider-operated) | While a phone is signed in with notifications on. |
| Google (Gemini API) | Kardo AI Cloud: phrasing answers that were already counted in code, understanding a question’s intent, enhanced video search over a few pictures, transcribing “Talk to the screen” recordings, and — when it is the configured provider — turning alert text into embeddings for search by meaning. | Question text, counted facts, alert text; a few small pictures for enhanced search where a site allows it; short voice recordings. | United States (provider-operated) | Optional. Only while the owner keeps Kardo AI Cloud switched on (Cloud & Privacy), and a site lets enhanced search show pictures. Everything on the KardoHub works without it. |
| Voyage AI | Embeddings: turning an alert’s title and summary into numbers so Search Moments can find alerts by meaning, when it is the configured embeddings provider. | Alert title and summary text only. No pictures. | United States (provider-operated) | Optional. Only while the owner keeps the semantic index of alerts switched on; switching it off deletes the index. |
[Owner: confirm each provider’s processing location and data-processing terms; “provider-operated” means Kardo has not chosen a region for that provider.] Where a camera, reader or phone from another maker connects through a KardoHub, that maker’s own service (for example a reader’s video-intercom relay) is not a Kardo sub-processor and is governed by its own terms.
Last updated 29 September 2026. KardoVision is a product of Kardo.