Model Zero
mizuraGet Mizura

Privacy policy.

Mizura keeps your research on your device, retains nothing from your prompts or data by default, and never shares your prompts to improve the model unless you explicitly opt in.

Who develops Mizura

Mizura is developed by Model Zero for Technological Solutions. Company information is available at mzfortech.com.

This website

This site has no account system, contact form, advertising, or analytics scripts. It does not intentionally set cookies or browser storage. The hosting provider may process request information such as IP addresses and access logs to deliver and secure the site.

Locally run Mizura

The desktop application runs an analysis engine on your computer and stores sessions, datasets, and outputs in a local data folder on your device. Analysis code runs on your device in a sandbox. Local execution does not mean every interaction stays on your device. Prompts and analysis context are sent to Model Zero, the model behind Mizura, to generate responses.

Zero data retention

Mizura is built on zero data retention (ZDR). We retain nothing from your prompts, datasets, or sessions, and we do not use them to train models. What is processed to generate a response is not kept afterwards.

Optional sharing for model improvement

You can help improve the model by opting in to share your prompts. Opt-in is off by default. Nothing is shared unless you turn it on. When it is on, the prompts you share may be retained and used to improve the model. You can turn it off at any time. Turning it off stops further sharing.

Feedback ratings

Thumbs up and thumbs down on a response record your rating of that response. Ratings show us what the model did right or wrong, and we use them as feedback on model quality. A rating never shares your prompts or data by itself. Sharing prompt content requires the separate opt-in above.

External services

Literature and web searches contact external services. Optional data connectors (GitHub, Google Drive, Zenodo) communicate with their respective providers when you connect them. Searches are screened against identifier-like terms from your data before anything is sent, so what is in your dataset stays out of your searches.

Exports and deletion

Reports and replication packages may include data, code, and execution records. Inspect exports before sharing. Deleting a local session or the application removes your local copies. It does not delete copies you exported elsewhere.

Questions and requests

Use the company's published contact channels at mzfortech.com. Tell us what you need and we will respond.