Privacy Statement Smardly Voice

This is a translation. In case of any difference, the Dutch version prevails.

Last updated: 6 November 2026

In one paragraph

Smardly Voice works entirely on your own device. Your recordings, your transcripts and your summaries do not go to us. They cannot, because the app makes no connection to Smardly. We can therefore not view your conversations, not store them and not provide them to others.

Who is responsible

Smardly B.V.
Overhoven 5, 6136 EA Sittard, the Netherlands
Chamber of Commerce 80435602
[email protected] — +31 85 743 00 23

What data we receive from you

None. Not your recordings, not your transcripts, not your summaries, not your name, not your email address, not your device details, not your app version and no usage statistics.

The app contains no analytics, no telemetry, no crash reporting to us, no advertising and no third-party trackers. There is no exception: the app never contacts us.

Where your data is stored

On your device. If you have iCloud synchronisation turned on, also in your own iCloud account, where Apple is the processor and we have no access.

If you do not want recordings to go to iCloud, turn off iCloud for this app in your device's settings. If your device is managed by your employer, that organisation can arrange this centrally.

Speech recognition and summarising

Both happen on your device, with models that run on the device. No audio and no text goes to Smardly, and none to an AI service of a third party. We do not use your conversations to train models — we do not have them.

Microphone and permissions

The app asks for permission to use the microphone, to record your conversations. The sound is kept on your device and processed there; it does not go to Smardly. For Apple's older speech recognition, which the app only uses when the other engines cannot, iOS also asks for permission for speech recognition. That processing also happens on your device. You can withdraw both permissions at any time in your device's settings. Without the microphone you cannot record.

Downloading models

The models for speech recognition and speaker recognition are not part of the app. Together they are about 475 MB. They arrive on your device once and stay there. After that the app works without a connection. If a model changes, the system fetches the new one when the app is updated.

In doing so, your device does connect to a party other than Smardly. These are files coming in; nothing of yours goes out, except what every internet request sends along, such as your device's IP address.

• Apple. Normally the models arrive as an asset pack from Apple (Background Assets, hosted by Apple). The system fetches them itself when you install or update the app. Apple's dictation models, which the app can use as a safety net, are also fetched from Apple by iOS itself.
• Hugging Face. If fetching through Apple does not work, for example in a version not installed through TestFlight or the App Store, the app fetches the models from Hugging Face. The app sends nothing along itself: no account, no device ID, no recording and no text.

This does not change the sentence that the app makes no connection to Smardly.

If fetching fails, the app tries the transcription with Apple's models on your device. If that fails too, the recording is kept and you get a message. With "Try again" you can try again later.

If your employer manages the app

If you get the app through your employer, that organisation can use its management software to pass settings to the app, for example that you may use Smardly Pro. That setting is on your own device. Nothing goes to Smardly for this, and we receive nothing about who uses the app.

Purchases in the App Store

If you buy Smardly Pro in the app, that purchase goes through Apple. Apple is your contract party for the payment; we receive no payment details from Apple. Apple's privacy policy applies to that purchase.

Sharing with others

You can email or share a report yourself. That goes from your device to your recipient, through your own mail app. It does not pass through us.

Sub-processors

For Smardly Voice: none. Smardly passes no data to third parties, and the sub-processors on our platform statement do not apply to this app. Your device does fetch the speech models from Apple and, as a fallback, from Hugging Face (see "Downloading models"). Smardly gives them no data; they are not sub-processors of Smardly.

Your rights

You have the right of access, rectification, erasure, restriction, objection and data portability. Because we have nothing of your recordings, we cannot view or delete them for you either — you do that yourself in the app. For questions you can email us at [email protected]. You can lodge a complaint with the Dutch Data Protection Authority (Autoriteit Persoonsgegevens).

Children

The app is not aimed at children under 16.

Changes

We change this statement when the app changes. If a processing activity is added, it will appear here before that processing goes live, and we will announce it in the app.

Terms and conditions
Privacy statement Smardly Cards
Data processing agreement Smardly Team
Security of Smardly Voice