Drive provider filters from the archive dataset

media-scrape/src/filter-presets.ts

@@ -48,3 +48,23 @@ export function resolutionCondition(preset: string): string | null {
             return null;
     }
 }
+
+/**
+ * The providers that dominate the archive, for the "Do not include" shortcuts.
+ * Aggregated 2026-09-17 over media-scrape/db (`provider` column, distinct URLs; discord folds in
+ * cdn.discordapp.com and media.discordapp.net). Matching is a case-insensitive substring on
+ * `provider`, so "twitter" also covers fxtwitter.com. The 4chan hosts stay out: the default URL
+ * host exclusion already hides them.
+ */
+export const PROVIDER_FILTERS = [
+    "discord",
+    "youtube",
+    "tenor",
+    "twitch",
+    "twitter",
+    "steam",
+    "medal",
+    "sketchfab",
+    "gfycat",
+    "imgur"
+] as const;

media-scrape/src/index.ts

@@ -10,6 +10,7 @@ export { isNearLimit, planRetention, type RetentionPlan, type RetentionScrape }
 export { isoFromSnowflake, snowflakeFromIso } from "./snowflake.js";
 export {
     ASPECT_PRESETS,
+    PROVIDER_FILTERS,
     RESOLUTION_PRESETS,
     aspectCondition,
     resolutionCondition,