--- title: Content Advisories & Safe Browsing description: How content ratings appear to visitors — inline advisories, blur-until-click, the sensitive-content statement, child-safe browsing, and the report queue. section: guides order: 13 updated: 2026-07-17 verified: 2026-07-17 related: [guides/content-screening, guides/cultural-sensitivity, guides/collections-and-permissions, guides/working-with-artifacts] features: [content-warnings] --- # Content Advisories & Safe Browsing This is the visitor-facing side of [content screening](/docs/en-US/guides/content-screening): how a content rating turns into what a visitor actually sees, and the institution settings that control it. The guiding principle is the museum one — **label and contextualize; almost never gate.** Warnings appear on flagged items; clean items show nothing. You configure all of this from **Admin → Site Settings → Content Safety**. ## What visitors see on a flagged item Two in-place treatments appear on the artifact viewing page, driven by the item's [content rating](/docs/en-US/guides/content-screening#how-screening-works): - **An advisory banner** sits above the item whenever it carries warnings. The banner is tinted by severity (amber for caution, red for restricted) and names *why* the material is preserved in a provenance-and-context voice — for example, that historical violence is kept for its research value. - **Blur-until-click** covers `restricted` media (nudity, explicit, or graphic imagery). The image stays in place, blurred, behind a short prompt such as "Contains nudity — click to view." One click reveals it. There is no full-page interstitial. `caution` items are never blurred — they render normally with the banner. The banner reads the warning *codes*, not just the rating, so a historical slur, a battlefield photograph, and a medical illustration each get appropriate wording. ### Reviewed vs. provisional warnings By default, only warnings a curator has approved appear to visitors. AI-suggested warnings stay in the [screening review queue](/docs/en-US/guides/content-screening#review-whats-flagged) until someone approves them. If you turn on **Show unreviewed AI warnings immediately**, AI warnings appear right away with an added line: *"This note was generated automatically and has not yet been reviewed by staff."* This is the honest-about-incompleteness posture — useful for large unreviewed backlogs, at the cost of occasional false positives. ## The sensitive-content statement Most archives publish a standing statement explaining why their collections contain offensive or distressing material. Preservated provides one with sensible default text, reachable from **a dismissible banner** on browse and collection landing pages. A visitor can dismiss it; it stays dismissed (remembered in their browser, and across devices when logged in). Editing the statement text re-shows it once. The banner is on by default. To change the wording, edit **Statement text** in Site Settings — leave it empty to use the platform default (shown as placeholder). To remove the statement page's only entry point, turn off **Show the dismissible landing banner**. ## Child-safe (filtered) browsing Filtered browsing — a "family-friendly" or kiosk mode — excludes `restricted` and unscreened items from discovery. Set **Filtered browsing** in Site Settings: - **Off** (default) — no filtering; restricted items appear in normal browsing (blurred) and unscreened items are treated as general. - **Optional** — visitors get a "Family-friendly view" toggle on the browse page and can switch it on for themselves. Their choice is remembered. - **Enforced** — the whole public site is always filtered. :::info Filtering applies everywhere a visitor discovers content — browse, search, facet counts, collection covers, related items, and visual-similarity results — not just the main grid. A deep link to a filtered-out item returns "not found" in filtered mode. Administrators are never filtered. ::: Unscreened items (those that haven't been through screening yet) are excluded from filtered browsing by default, since their rating isn't known. Run **Screen unscreened images** from the [screening queue](/docs/en-US/guides/content-screening#screen-existing-images) to clear the backlog. ## Visitor reports Every artifact viewing page carries a **Report content** item in the Share menu (the "More share options" button next to Share). A visitor can flag an item and add an optional note. Reports land in **Admin → Operations → Reports** for a curator to triage — mark reviewed or dismiss. Reporting does not change the item automatically; it is a signal for staff, who weigh potential harm against accurate preservation of the historical record. :::caution This curatorial report queue is **not** the path for suspected illegal content. Material that may be illegal is handled separately, through the "Flagged for human review" list in [Content Screening](/docs/en-US/guides/content-screening#review-whats-flagged) — report, don't investigate. ::: ## Settings summary | Setting | Default | Effect | | --- | --- | --- | | Filtered browsing | Off | Off / Optional (visitor toggle) / Enforced | | Show unreviewed AI warnings immediately | Off | Display AI warnings before curator approval, marked provisional | | Show the "sensitive content" statement | On | Report content menu item (artifact page) | | Show the dismissible landing banner | On | One-time banner on landing pages | | Statement text | Platform default | Markdown override of the statement body | Per-item warnings always appear on flagged items regardless of these settings — they label, they don't gate.