From ca5f0a91c0009435d2a81cf7cc0606c92bfb0dda Mon Sep 17 00:00:00 2001
From: Vuong Hoang <vh@phasefinal.com>
Date: Fri, 18 Sep 2026 12:42:51 -0700
Subject: [PATCH] searxng: sync settings with live (captcha-era engine set)

---
 stacks/searxng/conf/searxng-settings.yml | 32 ++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/stacks/searxng/conf/searxng-settings.yml b/stacks/searxng/conf/searxng-settings.yml
index 03f3208..04fd972 100644
--- a/stacks/searxng/conf/searxng-settings.yml
+++ b/stacks/searxng/conf/searxng-settings.yml
@@ -7,6 +7,11 @@
 #   residential SOCKS5 egress proxy on nh3-dev for yt-dlp. Running the search
 #   aggregator from a residential-egress site removes the problem at the source
 #   rather than proxying around it.
+#   UPDATE 2026-09-18: residential alone no longer sufficient — DDG + Startpage
+#   now captcha even here. Tier A fix: scraped families removed, API-backed and
+#   aggregator-tolerant engines kept. If Brave results are wanted, get a free
+#   Brave Search API key and set `api_key` on the `braveapi` engine (inactive,
+#   takes priority=0 when keyed) instead of returning to scraping.
 
 use_default_settings:
   engines:
@@ -21,6 +26,33 @@ use_default_settings:
       # The name is "karmasearch videos".
       - karmasearch
       - karmasearch videos
+      # Tier A (v3, 2026-09-18): scraped engine families that CAPTCHA/challenge
+      # this egress IP. Measured 3/3 queries: duckduckgo -> CAPTCHA, startpage
+      # -> Suspended: CAPTCHA, brave -> Suspended: too many requests. The whole
+      # family goes: images/videos/news siblings share the same scraper and
+      # network. API-backed engines never see captchas.
+      - brave
+      - brave.images
+      - brave.videos
+      - brave.news
+      - duckduckgo
+      - duckduckgo images
+      - duckduckgo videos
+      - duckduckgo news
+      - startpage
+      - startpage news
+      - startpage images
+      # Wikidata (an API engine, not scraped): Wikimedia answers this egress
+      # with 403 (measured direct from the host, 6/6 searxng probes + a
+      # manual sparql probe). Only feeds HTML infoboxes; keeping it meant a
+      # permanent 'Suspended' entry in every response.
+      - wikidata
+
+# Breadth note (v2 experiment): mojeek was re-enabled and is ALSO 403'd from
+# this egress IP (3/3 probes + direct 403 from the host, with and without a
+# browser UA). Re-adding breadth now needs API keys, not scraping: braveapi
+# (free Brave Search API tier) and/or marginalia (free key) — both already
+# present in the defaults, awaiting an api_key.
 
 general:
   instance_name: "SearXNG"