Your work.
Know where it goes.
A local desktop workspace and an AI endpoint are different things. Understanding that boundary is the starting point for control.
The workspace
Project folders, conversation context and app configuration.
The endpoint
The selected model server processes the context sent to it.
Control is a practice.
Not a slogan.
The screenshots describe a local-first desktop app. They do not establish the retention or logging policy of every service you connect.
Does local mean offline?
Not necessarily. A desktop app runs locally, but a remote model endpoint still receives requests. Offline work requires a suitable local model and tools that do not need network access.
Who can receive project context?
The configured model endpoint and any external tools involved in a task may receive relevant context. Review the provider, tool permissions and requested data before connecting sensitive projects.
Are provider logs controlled by Uncensored Coding?
Do not assume so. Retention, logging and training policies are determined by the service you choose. Check its terms and configuration separately.
How should I handle secrets?
Use narrowly scoped credentials. Avoid placing unnecessary secrets in prompts or files shared with a model. Review tool access and revoke unused keys.
Is this a formal privacy policy?
This page explains the workspace and endpoint boundaries. It is not a substitute for a published legal policy or an independent security assessment.
Your tools.
Your decisions.
Choose connections carefully.
Share only the context the task needs.
Keep consequential actions deliberate.