Add degoog as the search provider, and make it the default

The operator already runs it, so this is the one provider that costs nothing to adopt: no key,
no card, no third party, and no new container. It aggregates several engines and returns them
merged, and its /api/search reply is already the shape everything downstream expects — title,
url, snippet, with content as a fallback for the roughly one result in six that has no snippet.

It is also now the only provider whose response has been read from a live service rather than
from documentation, which is why it is the default.

Deployment/host.conf.template

@@ -585,6 +585,7 @@
 # Per-host because one is an address on this network and the other is a credential. Only the
 # General Chat profile can use these — it is the profile that cannot change anything, which is
 # why it is the one allowed to look outside. Off until AI_WEB_SEARCH_ENABLED says otherwise.
+    HOSTN_DEGOOG_URL=""                          # e.g. http://localhost:4444 — self-hosted, no key, /api/search
     HOSTN_SEARXNG_URL=""                         # e.g. http://localhost:8888 — needs format: [json] in its settings.yml
     HOSTN_WEB_SEARCH_API_KEY=""                  # brave or tavily; unused when the provider is searxng
     HOSTN_OLLAMA_EMBED_MODEL="nomic-embed-text"  # embeddings — the generation model cannot embed

Deployment/master.conf.template

@@ -1837,13 +1837,15 @@
 # when it is ticked. A question about this machine hands off to the assistant before the search
 # would run, so it never reaches the internet even with the box ticked.
 #
-# Provider: searxng | brave | tavily
+# Provider: degoog | searxng | brave | tavily
+#   degoog   self-hosted, no key — set HOST*_DEGOOG_URL. Aggregates several engines and returns
+#            them merged. The default, and the only one verified against a live service here
 #   searxng  self-hosted, no key, no third party — set HOST*_SEARXNG_URL, and enable format: [json]
 #            in its own settings.yml, which the default image ships with off
 #   brave    HOST*_WEB_SEARCH_API_KEY, free tier available
 #   tavily   HOST*_WEB_SEARCH_API_KEY
     AI_WEB_SEARCH_ENABLED=false
-    AI_WEB_SEARCH_PROVIDER=searxng
+    AI_WEB_SEARCH_PROVIDER=degoog
     AI_WEB_SEARCH_RESULTS=4
     AI_WEB_SEARCH_TIMEOUT=6
 

Plugin/unraid/include/ai_web.php

@@ -5,11 +5,16 @@
 //   their URLs, for the worker to put in front of the model as context it can cite.
 //
 // OPERATIONAL MODEL
-//   One provider at a time, named by AI_WEB_SEARCH_PROVIDER. Three are supported because the
+//   One provider at a time, named by AI_WEB_SEARCH_PROVIDER. Four are supported because the right
-//   right answer depends on what the operator is willing to run: a self-hosted SearXNG needs no
+//   answer depends on what the operator already runs: degoog and SearXNG are self-hosted and need
-//   key and no third party, and a hosted API needs no container. They differ only in how the
+//   no key and no third party, while brave and tavily need no container. They differ only in how
-//   request is shaped and where the fields sit in the reply, so everything after parsing is one
+//   the request is shaped and where the fields sit in the reply, so everything after parsing is
-//   code path.
+//   one code path.
+//
+//   degoog is the default and the one verified here — it was already running on this host, and it
+//   aggregates several engines and returns them merged. DuckDuckGo is deliberately absent: both
+//   its scraping endpoints answer a challenge page from this network, and its keyless API returns
+//   nothing at all for a real question.
 //
 // DESIGN PRINCIPLES
 //   Off by default, and not because it is dangerous.
@@ -43,7 +48,7 @@
 // ═══════════════════════════════════════════════════════════════════════════════════════════════
 require_once __DIR__ . '/ai.php';
 
-const VV_AI_WEB_PROVIDERS   = ['searxng', 'brave', 'tavily'];
+const VV_AI_WEB_PROVIDERS   = ['degoog', 'searxng', 'brave', 'tavily'];
 const VV_AI_WEB_MAX_QUERY   = 300;   // characters sent to the provider
 const VV_AI_WEB_MAX_SNIPPET = 400;   // characters kept from each result
 
@@ -53,8 +58,8 @@ function vv_ai_web_enabled(): bool {
 }
 
 function vv_ai_web_provider(): string {
-    $p = strtolower(trim((string)(vv_conf_vars()['AI_WEB_SEARCH_PROVIDER'] ?? 'searxng')));
+    $p = strtolower(trim((string)(vv_conf_vars()['AI_WEB_SEARCH_PROVIDER'] ?? 'degoog')));
-    return in_array($p, VV_AI_WEB_PROVIDERS, true) ? $p : 'searxng';
+    return in_array($p, VV_AI_WEB_PROVIDERS, true) ? $p : 'degoog';
 }
 
 function vv_ai_web_results_max(): int {
@@ -71,6 +76,15 @@ function vv_ai_web_searx_url(): string {
     return rtrim(trim((string)(vv_conf_vars()[strtoupper(vv_detect_host()) . '_SEARXNG_URL'] ?? '')), '/');
 }
 
+function vv_ai_web_degoog_url(): string {
+    return rtrim(trim((string)(vv_conf_vars()[strtoupper(vv_detect_host()) . '_DEGOOG_URL'] ?? '')), '/');
+}
+
+// Which providers need an address on this network rather than a key. Kept as a list because the
+// readiness check and the error message both need the same answer and disagreeing about it is how
+// a feature reports itself ready and then does nothing.
+const VV_AI_WEB_SELF_HOSTED = ['degoog', 'searxng'];
+
 function vv_ai_web_api_key(): string {
     return trim((string)(vv_conf_vars()[strtoupper(vv_detect_host()) . '_WEB_SEARCH_API_KEY'] ?? ''));
 }
@@ -82,11 +96,16 @@ function vv_ai_web_ready(): array {
         return ['ok' => false, 'error' => 'AI_WEB_SEARCH_ENABLED is false'];
     }
     $p    = vv_ai_web_provider();
-    if ($p === 'searxng' && vv_ai_web_searx_url() === '') {
+    $host = strtoupper(vv_detect_host());
-        return ['ok' => false, 'error' => strtoupper(vv_detect_host()) . '_SEARXNG_URL is empty'];
+
+    if ($p === 'degoog' && vv_ai_web_degoog_url() === '') {
+        return ['ok' => false, 'error' => $host . '_DEGOOG_URL is empty'];
     }
-    if ($p !== 'searxng' && vv_ai_web_api_key() === '') {
+    if ($p === 'searxng' && vv_ai_web_searx_url() === '') {
-        return ['ok' => false, 'error' => strtoupper(vv_detect_host()) . '_WEB_SEARCH_API_KEY is empty'];
+        return ['ok' => false, 'error' => $host . '_SEARXNG_URL is empty'];
+    }
+    if (!in_array($p, VV_AI_WEB_SELF_HOSTED, true) && vv_ai_web_api_key() === '') {
+        return ['ok' => false, 'error' => $host . '_WEB_SEARCH_API_KEY is empty'];
     }
     return ['ok' => true, 'provider' => $p];
 }
@@ -144,6 +163,25 @@ function vv_ai_web_normalise(array $rows): array {
 // a shape that changes underneath produces zero results rather than wrong ones, because every
 // field is read by name and anything missing is dropped by vv_ai_web_normalise().
 
+// Degoog — self-hosted, no key, and the only provider here whose response has been read from a
+// live service rather than from documentation. It aggregates several engines and returns them
+// already merged, so the ranking is its own.
+//
+// snippet before content: it fills both, and around one result in six has neither. Those are kept
+// anyway — a title and a URL is still a citable source, and dropping them would silently shrink
+// the result set the operator asked for.
+function vv_ai_web_degoog(string $q): ?array {
+    $base = vv_ai_web_degoog_url();
+    if ($base === '') return null;
+    $d = vv_ai_web_http($base . '/api/search?q=' . rawurlencode($q));
+    if ($d === null) return null;
+    return array_map(fn($r) => ['title'   => $r['title'] ?? '',
+                                'url'     => $r['url'] ?? '',
+                                'snippet' => ($r['snippet'] ?? '') !== '' ? $r['snippet']
+                                                                          : ($r['content'] ?? '')],
+                     array_slice($d['results'] ?? [], 0, vv_ai_web_results_max()));
+}
+
 function vv_ai_web_searxng(string $q): ?array {
     $base = vv_ai_web_searx_url();
     if ($base === '') return null;
@@ -201,7 +239,8 @@ function vv_ai_web_search(string $question): array {
     $raw = match ($provider) {
         'brave'   => vv_ai_web_brave($q),
         'tavily'  => vv_ai_web_tavily($q),
-        default  => vv_ai_web_searxng($q),
+        'searxng' => vv_ai_web_searxng($q),
+        default   => vv_ai_web_degoog($q),
     };
 
     if ($raw === null) {