Omnet Nuvori Privacy Policy

On-device chat

Omnet Nuvori is a native Android app for general on-device chat. Messages you send and replies the app generates are processed locally on your device. After the model is on the device, chat runs there with no cloud inference fallback. Generated text can be wrong or outdated. Speed and quality depend on your device. Optional safety reporting is separate and is described below.

Data the app handles

The app stores conversation titles, timestamps, roles, messages, and in-progress replies, along with the language model, in ordinary Android private storage. This release does not add its own encryption on top of that storage.

Optional safety reporting

You can choose Report in the app. Report first shows a native disclosure. Then you tap Open report form, which contacts Google and opens a blank Google Form. The app does not prefill the form or attach chat automatically. You may copy a selected reply on your device, then paste and review the excerpt you choose to share.

The form has a required question, “What is wrong with the AI output?”, and an optional question, “AI output you choose to share (optional).” It does not collect an email address. You do not have to sign in, and you are not limited to one report. Form autosave and response summaries are off.

Opening the form contacts Google and sends ordinary connection information, such as IP address and request metadata. Text you type is processed by Google. If you click Submit, the report is sent to Omnet as a Google Forms response. We use reports to investigate safety. You do not have to provide an email address or name. Please do not include personal information you do not want to share.

Other conversations and on-device inference stay on your device. Submitted reports are stored in Google Forms that Omnet manages, separate from your phone. Deleting local chats or the model does not delete reports you already submitted.

Google’s handling of form and connection data is described in Google’s privacy policy. The reporting form is this Google Form.

Model download and import

Downloading the model needs a network connection and uses the Android INTERNET permission. The download is the official Hugging Face Qwen3 1.7B GGUF model, about 1.83 GB, and ordinary content-delivery network redirects are expected. You can also import an official copy of that model; the original file outside the app is left unchanged.

Hugging Face and any content-delivery network that serves the file receive ordinary request information, such as IP address and request metadata. See Hugging Face’s privacy policy.

Storage and deletion

The app offers separate controls to remove conversation history and to remove the downloaded model. Removing the model does not remove conversations. You can also clear Android app storage or uninstall the app, which removes app-local data under ordinary Android platform behavior. Those local actions do not delete safety reports you already submitted.

Backup

Android cloud backup and device transfer are disabled for this release.

Accounts, ads, and purchases

This release has no in-app account, no cloud inference, no ads, and no in-app billing. It does not add an analytics or crash-reporting service. Optional safety reporting uses Google Forms when you choose Report, as described above.

Store, website, and support

Google Play, Android, and other platform services may process information under their own policies when the app is distributed or run on your device. That processing is separate from the local data handling described above. If you open this privacy page, it loads in an external browser as an ordinary website request to our hosting provider; the app does not contact this page automatically.

If you contact support, the message includes what you choose to send to support@omnetsystems.com, plus ordinary email metadata. Please do not send passwords or sensitive personal information.

Contact

Omnet Nuvori is an Omnet app. For privacy questions, report deletion requests, or other help, email support@omnetsystems.com. To help locate an anonymous report, include its approximate submission time and relevant details when requesting deletion.