From 8cb1faf8b4f4f20fd8957721f593fc21fb4acea7 Mon Sep 17 00:00:00 2001
From: Fisto <fisto@fisto.cccp>
Date: Thu, 17 Sep 2026 11:03:18 +0200
Subject: [PATCH] Queue the data-driven provider filter list

---
 TODO.md | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/TODO.md b/TODO.md
index fd271c0..b581183 100644
--- a/TODO.md
+++ b/TODO.md
@@ -62,6 +62,16 @@ Ordered backlog. `CURRENT.md` tracks the live session state.
     optional `VOX_CLEAN`, and a help line in the TTS panel. Requested in `TTS-DESIGN.md`; the
     wrapper side is ready and verified. Inline `/rate` and `/pitch` need no UI.
 
+## Media archive follow-ups
+
+1. **Provider exclusion list from real data.** The Advanced `Do not include → Sources` checkboxes
+   use a hand-picked list (Twitch, YouTube, Discord, Tenor, Giphy). Replace it with the providers
+   that actually dominate the archive: aggregate the `provider` column across the scraper
+   databases (`SELECT lower(provider), COUNT(*) FROM media WHERE provider IS NOT NULL GROUP BY 1
+   ORDER BY 2 DESC`), take the top ~8-10, turn that into the checkbox list, and unit-test that the
+   UI list matches the shared constant. Keep the source-kind boxes (embed/attachment/sticker)
+   separate from the provider boxes.
+
 ## Web UI follow-ups
 
 Ordered after the voice backend unless the owner asks first.