Privacy policy.

Last updated August 1, 2026

On-device processing

Watt Inference runs models directly on your device. Prompts, responses, and any content you generate are processed locally and are not transmitted to our servers or any third party as part of normal use.

Information we collect

We collect limited account and usage information — such as your email address if you request access, and basic app diagnostics — to operate the service, respond to support requests, and improve reliability. We do not collect the content of your on-device conversations.

How we use information

Information we collect is used to provide and maintain the app, communicate with you about access and updates, and understand aggregate, anonymized usage patterns. We do not sell personal information.

Open source weights

Model weights published on Hugging Face are public and governed by their own license terms, separate from this policy.

Your choices

You can request deletion of any account information we hold by contacting us. Because inference happens on-device, uninstalling the app removes any local model data.

Contact

Questions about this policy can be sent to [email protected].

This is placeholder policy text and should be reviewed by counsel before publishing.