Privacy Policy
OpenHead Robot Face for iOS and iPadOS · Effective 3 November 2026 (camera stream to a paired viewer added)
In short: the app collects no personal data. It has no accounts, analytics, advertising or tracking, and it sends nothing to us or to any third party. It only talks to devices on your own local network that you have paired with it.
Who we are
OpenHead is an open-source project published on GitHub at github.com/PatrikBichis/openhead. "We" in this policy means the developer of the OpenHead Robot Face app. Questions about this policy can be asked on the support page.
Data we collect
None. The app does not collect, store on our behalf, sell or share personal data, and it contains no third-party SDKs for analytics, crash reporting or advertising.
What the app does on your device and network
- Local network. The app runs a small server on your Wi-Fi so that a robot controller can connect, and announces itself with Bonjour (mDNS) so it can be found. iOS asks for permission to use the local network. Connections come only from devices on the same network.
- Pairing. A controller must type a 6-digit code shown on the screen. Each paired controller gets a random token, stored with its name and last-seen time in the device's Keychain. Nothing about it leaves the device. You can revoke any controller in the app's settings.
- Motion sensors. If a paired controller subscribes, the app streams the device's orientation, rotation and acceleration to that controller, on your local network only. This needs no permission and is not stored.
- Pictures, video and web pages. When a paired controller asks the app to show an image, a video or a web page, the app fetches it from the address the controller gives. That server sees the request as it would from any browser. The app does not keep the content after it is no longer shown.
- Logs and screenshots. The app keeps a short log of its own activity in memory (cleared when the app quits) for troubleshooting. With dev mode switched on in settings, a paired controller can read that log and take screenshots of what is on screen. Dev mode is off by default.
- Settings. Your settings (name, port, theme, mounting, dev mode) are stored on the device only.
- Camera. The camera is only used when a paired controller switches on a camera mode, and iOS asks for your permission the first time. In capture mode the front camera shows what the head sees on its own screen and finds faces; in sensing mode ARKit uses the cameras to track where the head is in the room and to follow one face. Camera images are processed on the device and are never stored. They leave the phone in one case only: a paired controller can ask for a small, low-frame-rate video stream of the camera (for example the cockpit on your computer); it goes only to that controller on your local network, and the head shows “Camera shared” on its screen the whole time. Otherwise what paired controllers receive are positions and angles: where faces are (direction, size, distance) and where the head is. Faces are not recognised or identified, and no face data is kept. With dev mode switched on, a paired controller can take a screenshot of the screen, which then includes the camera view.
- Location. Only when a paired controller subscribes to the
gpsstream does the app ask iOS for your location ("While Using the App"), and iOS asks for your permission the first time. The location is sent only to the paired controllers on your local network that asked for it, while they are connected. It is not stored, not sent to us and not used for anything else. With location allowed, the compass also reports true north. - Compass, magnetometer and barometer. The heading on the camera view comes from the compass. If a paired controller subscribes to the
envstream, the app sends it the compass heading, the magnetic field and the air pressure and altitude change from the barometer, on your local network only. The barometer needs Motion & Fitness permission, which iOS asks for the first time. None of this is stored. - Speaking. When a paired controller asks the head to speak (
speech.say), the app turns the text into speech with Apple's voices on the device and plays it on the phone's speaker or a connected speaker. The text is not stored. - Microphone and speech recognition. Only when a paired controller asks the head to listen (
speech.listen) does the app use the microphone, and iOS asks for your permission the first time, for the microphone and for speech recognition. What you say is turned into text with Apple's speech recognition, on the device when the language supports it; otherwise Apple processes the audio under Apple's privacy policy. Audio is not recorded or stored. The text goes only to the paired controllers on your local network that asked for it. The app stops listening when it is told to, and when the last controller disconnects; while it listens, iOS shows the microphone indicator. The controller decides what to do with the text, for example send it to an AI service of your choice.
Children
The app collects no data from anyone, including children.
Data retention and deletion
Everything the app stores stays on your device. Deleting the app removes its settings and pairing tokens; revoking a controller in settings removes its token immediately.
Security
Only paired controllers can change what the app shows; five wrong pairing codes lock pairing for five minutes. Traffic on the local network is not encrypted in this version, so use the app on a network you trust.
Changes to this policy
If this policy changes, the new version is published on this page with a new effective date, and the history is kept in the project's public Git repository.
Contact
Open an issue on GitHub or see the support page.