Last updated: June 2026

<aside> 🛡️ Sanctum is a focus-support browser extension for YouTube. Its purpose is to help users notice and interrupt repetitive browsing patterns such as feed checking, Shorts scrolling, refresh loops, and impulsive video selection.

</aside>

1. Data processed locally

Sanctum’s core detection and decision logic run in the user’s browser.

Sanctum may store the following data locally in the user’s browser:

• User settings such as language preference, focus mode, feedback position, excluded domains, and opt-in preference
• Local feedback records
• Local media cache
• Recent playback history
• Local extension state needed for focus-support features

This local data is used only to provide Sanctum’s focus-support features.

2. Optional improvement data

Feedback and Improvement Data (Stable Channel Disclosure):

• When a user explicitly presses Send feedback in the Anonymous Feedback form, the entered free-form text is transmitted to Firebase. It is used only for product improvement and error review, and users are advised not to include personal or sensitive information.
• The popup feedback payload is limited to type, feedback_text, created_at, app_version, locale, and source. It does not include browsing history, the current URL, video titles, page content, account information, or any IP address collected by the extension.
• The improvement-data opt-in controls automatic structured events. If enabled, only a structured summary of 13 specific technical fields is transmitted to Firebase: schema_version, client_version, event_type, rule_id, rule_category, intervention_type, user_action, post_action_pattern, time_to_next_reentry_bucket, time_to_escape_bucket, feedback_label, collection_mode, and created_at.
• The uninstall survey is entirely optional and distinct from the in-extension feedback. The survey URL includes only the source, extension version, and browser locale parameters; no free-form text is automatically submitted.

If the user enables improvement data sharing, Sanctum may send minimal non-identifying summary events for rule improvement. These events may include:

• Rule ID
• Rule category
• Intervention type
• User action result
• Feedback label
• Time bucket values