diff --git a/changelog.d/features/12167-universal-context-handoff-flag.md b/changelog.d/features/12167-universal-context-handoff-flag.md new file mode 100644 index 000000000000..4355bc86aa28 --- /dev/null +++ b/changelog.d/features/12167-universal-context-handoff-flag.md @@ -0,0 +1 @@ +- Add a runtime feature flag to disable universal context handoffs globally without changing the default behavior. diff --git a/open-sse/services/contextHandoff.ts b/open-sse/services/contextHandoff.ts index 9ac2ffdd14ed..bd169be60675 100644 --- a/open-sse/services/contextHandoff.ts +++ b/open-sse/services/contextHandoff.ts @@ -7,6 +7,7 @@ import { } from "../../src/lib/db/contextHandoffs.ts"; import { estimateTokens } from "./contextManager.ts"; import { stripMarkdownCodeFence } from "../utils/aiSdkCompat.ts"; +import { isFeatureFlagEnabled } from "../../src/shared/utils/featureFlags.ts"; export const HANDOFF_WARNING_THRESHOLD = 0.85; export const HANDOFF_EXHAUSTION_THRESHOLD = 0.95; @@ -139,7 +140,9 @@ export function resolveUniversalHandoffConfig( triggerRaw === "always" || triggerRaw === "on-error" ? triggerRaw : "on-switch"; return { - enabled: getBool("enabled", DEFAULT_UNIVERSAL_HANDOFF_CONFIG.enabled), + enabled: + isFeatureFlagEnabled("UNIVERSAL_CONTEXT_HANDOFF_ENABLED") && + getBool("enabled", DEFAULT_UNIVERSAL_HANDOFF_CONFIG.enabled), trigger, providerAllowlist: getStringArray( "providerAllowlist", diff --git a/src/i18n/messages/ar.json b/src/i18n/messages/ar.json index 508f5c9cb10b..4844c7c3b56b 100644 --- a/src/i18n/messages/ar.json +++ b/src/i18n/messages/ar.json @@ -987,6 +987,7 @@ "disabled": "معطل", "featureFlagOmnirouteEmergencyFallbackDescription": "توجيه الطلبات التي استنفدت الميزانية إلى موفر/نموذج الاحتياط المجاني للطوارئ.", "featureFlagArenaEloSyncEnabledDescription": "تمكين المزامنة الدورية لتصنيف ELO للوحة صدارة Arena AI لتصنيفات ذكاء النماذج.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "اعلن عن معرفات مرآة claude/<provider>/<model> على /v1/models حتى تظهر قائمة اكتشاف نماذج بوابة Claude Code نماذج غير Claude. تحذير: يؤدي إلى تكرار إدخالات الكتالوج لجميع العملاء عند تفعيله عالميًا.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "فعّل مسارات القبول الافتراضية التكيفية لكل مستأجر (tenant) لتوزيع المزودين (#9654): لم يعد انفجار حركة أحد المستأجرين يسبب خطأ 503 لمستأجر آخر. متغير البيئة OMNIROUTE_CHAT_VIRTUAL_LANES له الأولوية على هذا الإعداد في لوحة التحكم؛ تصبح التغييرات سارية بعد إعادة تشغيل الخادم.", diff --git a/src/i18n/messages/az.json b/src/i18n/messages/az.json index b44437a739b9..1fba41a708f8 100644 --- a/src/i18n/messages/az.json +++ b/src/i18n/messages/az.json @@ -987,6 +987,7 @@ "disabled": "Deaktiv", "featureFlagOmnirouteEmergencyFallbackDescription": "Büdcəsi tükənmiş sorğuları təcili pulsuz ehtiyat təminatçıya/modelə yönləndirin.", "featureFlagArenaEloSyncEnabledDescription": "Model intellekti reytinqləri üçün dövri Arena AI liderlər cədvəli ELO sinxronizasiyasını aktivləşdirin.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models üzərində claude/<provider>/<model> güzgü id-lərini reklam edin ki, Claude Code keçid modeli kəşfiyyatında qeyri-Claude modelləri siyahıya alsın. Diqqət: qlobal olaraq aktivləşdirildikdə bütün müştərilər üçün kataloq girişlərini ikiqat artırır.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Təchizatçı göndərişi üçün hər bir icarəçi (tenant) üzrə adaptiv virtual qəbul zolaqlarını aktivləşdirin (#9654): bir icarəçinin ani yükü artıq digərinə 503 qaytarmır. OMNIROUTE_CHAT_VIRTUAL_LANES mühit dəyişəni bu idarəetmə paneli ayarından üstündür; dəyişikliklər server yenidən işə salındıqda qüvvəyə minir.", diff --git a/src/i18n/messages/bg.json b/src/i18n/messages/bg.json index b89de00acee7..5368548a4dd2 100644 --- a/src/i18n/messages/bg.json +++ b/src/i18n/messages/bg.json @@ -987,6 +987,7 @@ "disabled": "Деактивирано", "featureFlagOmnirouteEmergencyFallbackDescription": "Маршрутизиране на заявки с изчерпан бюджет към аварийния безплатен резервен доставчик/модел.", "featureFlagArenaEloSyncEnabledDescription": "Активиране на периодична синхронизация на ELO от класацията на Arena AI за класиране на интелигентността на моделите.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Рекламирайте claude/<provider>/<model> mirror идентификатори на /v1/models, така че списъкът с модели на Claude Code gateway да включва неклаудови модели. Внимание: удвоява записите в каталога за всички клиенти, когато е активирано глобално.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Активирайте адаптивни виртуални ленти за допускане за всеки наемател (tenant) при изпращане към доставчици (#9654): скокът в натоварването на един наемател вече не връща 503 на друг. Променливата на средата OMNIROUTE_CHAT_VIRTUAL_LANES има предимство пред тази настройка в таблото; промените влизат в сила след рестартиране на сървъра.", diff --git a/src/i18n/messages/bn.json b/src/i18n/messages/bn.json index 26815e23b7b8..db95d7b2bffa 100644 --- a/src/i18n/messages/bn.json +++ b/src/i18n/messages/bn.json @@ -987,6 +987,7 @@ "disabled": "নিষ্ক্রিয়", "featureFlagOmnirouteEmergencyFallbackDescription": "বাজেট শেষ হয়ে যাওয়া অনুরোধগুলো জরুরি ফ্রি ফলব্যাক প্রোভাইডার/মডেলে রুট করুন।", "featureFlagArenaEloSyncEnabledDescription": "মডেল ইন্টেলিজেন্স র‍্যাঙ্কিংয়ের জন্য পর্যায়ক্রমিক Arena AI লিডারবোর্ড ELO সিঙ্ক সক্ষম করুন।", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models এ claude/<provider>/<model> মিরর আইডি বিজ্ঞাপন দিন যাতে Claude Code গেটওয়ে মডেল আবিষ্কার non-Claude মডেল তালিকাভুক্ত করে। সতর্কতা: এটি গ্লোবালি সক্ষম হলে সমস্ত ক্লায়েন্টের জন্য ক্যাটালগ এন্ট্রি দ্বিগুণ করে।", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "প্রোভাইডার ডিসপ্যাচের জন্য প্রতি-টেন্যান্ট অ্যাডাপ্টিভ ভার্চুয়াল অ্যাডমিশন লেন সক্ষম করুন (#9654): এক টেন্যান্টের বিস্ফোরণ আর অন্য টেন্যান্টে 503 ফেরায় না। OMNIROUTE_CHAT_VIRTUAL_LANES এনভায়রনমেন্ট ভেরিয়েবল এই ড্যাশবোর্ড সেটিংয়ের উপরে প্রাধান্য পায়; পরিবর্তনগুলি সার্ভার পুনরায় চালু হলে কার্যকর হয়।", diff --git a/src/i18n/messages/cs.json b/src/i18n/messages/cs.json index 29a078e0ff6a..e41706918835 100644 --- a/src/i18n/messages/cs.json +++ b/src/i18n/messages/cs.json @@ -987,6 +987,7 @@ "disabled": "Zakázáno", "featureFlagOmnirouteEmergencyFallbackDescription": "Směrovat požadavky s vyčerpaným rozpočtem na nouzového bezplatného záložního poskytovatele/model.", "featureFlagArenaEloSyncEnabledDescription": "Povolit periodickou synchronizaci ELO z žebříčku Arena AI pro hodnocení inteligence modelů.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Inzerujte claude/<provider>/<model> zrcadlové ID na /v1/models, aby seznam objevování modelů brány Claude Code zahrnoval modely, které nejsou Claude. Upozornění: při globálním povolení zdvojuje katalogové položky pro všechny klienty.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Povolte adaptivní virtuální vstupní pruhy pro každého tenanta při odesílání poskytovatelům (#9654): špička jednoho tenanta už nezpůsobí 503 u jiného. Proměnná prostředí OMNIROUTE_CHAT_VIRTUAL_LANES má přednost před tímto nastavením na řídicím panelu; změny se projeví po restartu serveru.", diff --git a/src/i18n/messages/da.json b/src/i18n/messages/da.json index 3ded2eedcdaa..f890027bf512 100644 --- a/src/i18n/messages/da.json +++ b/src/i18n/messages/da.json @@ -987,6 +987,7 @@ "disabled": "Deaktiveret", "featureFlagOmnirouteEmergencyFallbackDescription": "Diriger budgetudtømte anmodninger til den gratis nød-fallback-udbyder/-model.", "featureFlagArenaEloSyncEnabledDescription": "Aktivér periodisk ELO-synkronisering fra Arena AI-førertavlen til rangering af modelintelligens.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Reklamer claude/<provider>/<model> spejl-id'er på /v1/models, så Claude Code gateway modelopdagelse viser ikke-Claude modeller. Advarsel: fordobler katalogposter for alle klienter, når det er aktiveret globalt.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktivér adaptive virtuelle adgangsbaner pr. tenant til providerudlevering (#9654): en tenants burst giver ikke længere en anden 503. Miljøvariablen OMNIROUTE_CHAT_VIRTUAL_LANES har forrang over denne dashboard-indstilling; ændringer træder i kraft ved genstart af serveren.", diff --git a/src/i18n/messages/de.json b/src/i18n/messages/de.json index aabfda5f6906..593a3f79bb6e 100644 --- a/src/i18n/messages/de.json +++ b/src/i18n/messages/de.json @@ -987,6 +987,7 @@ "disabled": "Deaktiviert", "featureFlagOmnirouteEmergencyFallbackDescription": "Anfragen mit erschöpftem Budget an den kostenlosen Notfall-Fallback-Anbieter/das Notfall-Fallback-Modell weiterleiten.", "featureFlagArenaEloSyncEnabledDescription": "Periodischen ELO-Abgleich der Arena AI-Bestenliste für Modell-Intelligenz-Rankings aktivieren.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Bewerben Sie claude/<provider>/<model> Spiegel-IDs auf /v1/models, damit die Claude Code-Gateway-Modellentdeckung Nicht-Claude-Modelle auflistet. Warnung: Verdoppelt Katalogeinträge für alle Clients, wenn global aktiviert.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktivieren Sie adaptive virtuelle Zulassungsspuren pro Tenant für die Provider-Zustellung (#9654): Ein Burst eines Tenants führt nicht mehr zu 503 bei einem anderen. Die Umgebungsvariable OMNIROUTE_CHAT_VIRTUAL_LANES hat Vorrang vor dieser Dashboard-Einstellung; Änderungen werden erst nach einem Serverneustart wirksam.", diff --git a/src/i18n/messages/en.json b/src/i18n/messages/en.json index cea92f3b07c9..15ed1f477e37 100644 --- a/src/i18n/messages/en.json +++ b/src/i18n/messages/en.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Advertise claude/<provider>/<model> mirror ids on /v1/models so Claude Code gateway model discovery lists non-Claude models. Warning: doubles catalog entries for all clients when enabled globally.", "featureFlagNoThinkingAliasEnabledDescription": "Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Enable per-tenant adaptive virtual admission lanes for provider dispatch (#9654): one tenant's burst no longer 503s another. The OMNIROUTE_CHAT_VIRTUAL_LANES env var wins over this dashboard override; changes take effect at server restart.", diff --git a/src/i18n/messages/es.json b/src/i18n/messages/es.json index bfd97ed5a663..adcf438bd6ff 100644 --- a/src/i18n/messages/es.json +++ b/src/i18n/messages/es.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Anunciar los ids de espejo claude/<provider>/<model> en /v1/models para que la lista de descubrimiento de modelos del gateway de Claude Code incluya modelos que no son de Claude. Advertencia: duplica las entradas del catálogo para todos los clientes cuando se habilita globalmente.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Activa carriles de admisión virtuales adaptativos por tenant para el envío de proveedores (#9654): el pico de un tenant ya no devuelve 503 a otro. La variable de entorno OMNIROUTE_CHAT_VIRTUAL_LANES tiene prioridad sobre esta opción del panel; los cambios surten efecto al reiniciar el servidor.", diff --git a/src/i18n/messages/fa.json b/src/i18n/messages/fa.json index 0886a02010ec..24c71145b415 100644 --- a/src/i18n/messages/fa.json +++ b/src/i18n/messages/fa.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "آگهی شناسه‌های آینه claude/<provider>/<model> را در /v1/models به‌گونه‌ای تنظیم کنید که لیست کشف مدل‌های دروازه کد Claude شامل مدل‌های غیر Claude باشد. هشدار: در صورت فعال‌سازی جهانی، ورودی‌های کاتالوگ را برای تمام مشتریان دو برابر می‌کند.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "خط‌های پذیرش مجازی تطبیقی به‌ازای هر مستاجر (tenant) را برای ارسال به ارائه‌دهندگان فعال کنید (#9654): افزایش ناگهانی بار یک مستاجر دیگر خطای 503 را برای مستاجر دیگر ایجاد نمی‌کند. متغیر محیطی OMNIROUTE_CHAT_VIRTUAL_LANES بر این تنظیم داشبورد اولویت دارد؛ تغییرات پس از راه‌اندازی مجدد سرور اعمال می‌شوند.", diff --git a/src/i18n/messages/fi.json b/src/i18n/messages/fi.json index 1b73262a8a1a..8693a255ae01 100644 --- a/src/i18n/messages/fi.json +++ b/src/i18n/messages/fi.json @@ -987,6 +987,7 @@ "disabled": "Poistettu käytöstä", "featureFlagOmnirouteEmergencyFallbackDescription": "Reititä budjettinsa ylittäneet pyynnöt varalla olevalle ilmaiselle varatarjoajalle/-mallille.", "featureFlagArenaEloSyncEnabledDescription": "Ota käyttöön jaksottainen Arena AI -tulostaulukon ELO-synkronointi mallien älykkyysluokituksia varten.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Mainosta claude/<provider>/<model> peilid tunnuksia /v1/models, jotta Claude Code -portin mallin löytölistalla näkyvät ei-Claude-mallit. Varoitus: kaksinkertaistaa luettelo-merkinnät kaikille asiakkaille, kun se on otettu käyttöön globaalisti.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Ota käyttöön mukautuvat virtuaaliset sisäänottokaistat vuokraajaa (tenant) kohti palveluntarjoajien välitystä varten (#9654): yhden vuokraajan kuormapiikki ei enää aiheuta 503-virhettä toiselle. Ympäristömuuttuja OMNIROUTE_CHAT_VIRTUAL_LANES ohittaa tämän hallintapaneelin asetuksen; muutokset tulevat voimaan palvelimen uudelleenkäynnistyksessä.", diff --git a/src/i18n/messages/fr.json b/src/i18n/messages/fr.json index 816cbc0a29e4..99c1e783c7ba 100644 --- a/src/i18n/messages/fr.json +++ b/src/i18n/messages/fr.json @@ -987,6 +987,7 @@ "disabled": "Désactivé", "featureFlagOmnirouteEmergencyFallbackDescription": "Router les requêtes ayant épuisé leur budget vers le fournisseur/modèle de secours gratuit d'urgence.", "featureFlagArenaEloSyncEnabledDescription": "Activer la synchronisation périodique de l'ELO du classement Arena AI pour les classements d'intelligence des modèles.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Afficher les identifiants miroir claude/<provider>/<model> dans /v1/models afin que la découverte de modèles de la passerelle Claude Code répertorie les modèles non-Claude. Attention : cette option double les entrées du catalogue pour tous les clients lorsqu'elle est activée globalement.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Activez des voies d'admission virtuelles adaptatives par tenant pour la répartition des fournisseurs (#9654) : le pic d'un tenant ne renvoie plus 503 à un autre. La variable d'environnement OMNIROUTE_CHAT_VIRTUAL_LANES prime sur ce réglage du tableau de bord ; les modifications prennent effet au redémarrage du serveur.", diff --git a/src/i18n/messages/gu.json b/src/i18n/messages/gu.json index ec483fcfc03b..34f7e442d67b 100644 --- a/src/i18n/messages/gu.json +++ b/src/i18n/messages/gu.json @@ -987,6 +987,7 @@ "disabled": "નિષ્ક્રિય કરેલ", "featureFlagOmnirouteEmergencyFallbackDescription": "બજેટ-સમાપ્ત વિનંતીઓને કટોકટીના મફત ફોલબેક પ્રદાતા/મોડેલ પર રૂટ કરો.", "featureFlagArenaEloSyncEnabledDescription": "મોડેલ ઇન્ટેલિજન્સ રેન્કિંગ માટે સમયાંતરે Arena AI લીડરબોર્ડ ELO સિંક સક્ષમ કરો.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models પર claude/<provider>/<model> મિરર આઈડીઓનું જાહેરાત કરો જેથી Claude Code ગેટવે મોડલ શોધી કાઢે છે non-Claude મોડલ. ચેતવણી: જ્યારે વૈશ્વિક રીતે સક્રિય કરવામાં આવે ત્યારે તમામ ક્લાયન્ટ માટે કૅટલોગ એન્ટ્રીઓ ડબલ કરે છે.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "પ્રોવાઇડર ડિસ્પેચ માટે પ્રતિ-ટેનન્ટ અનુકૂલનશીલ વર્ચ્યુઅલ એડમિશન લેન સક્ષમ કરો (#9654): એક ટેનન્ટનો બર્સ્ટ હવે બીજા ટેનન્ટને 503 આપતો નથી. OMNIROUTE_CHAT_VIRTUAL_LANES એન્વાયર્નમેન્ટ વેરિયેબલ આ ડેશબોર્ડ સેટિંગ કરતાં વધુ પ્રાધાન્ય ધરાવે છે; ફેરફારો સર્વર પુનઃપ્રારંભ પર અસરકારક થાય છે.", diff --git a/src/i18n/messages/he.json b/src/i18n/messages/he.json index 0d7e77e2c63a..17eecc1ad574 100644 --- a/src/i18n/messages/he.json +++ b/src/i18n/messages/he.json @@ -987,6 +987,7 @@ "disabled": "מושבת", "featureFlagOmnirouteEmergencyFallbackDescription": "ניתוב בקשות שחרגו מהתקציב לספק/מודל גיבוי חינמי לשעת חירום.", "featureFlagArenaEloSyncEnabledDescription": "הפעלת סנכרון ELO תקופתי מלוח המובילים של Arena AI עבור דירוגי אינטליגנציית מודלים.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "פרסם את מזהי המראה של claude/<provider>/<model> ב-/v1/models כך שרשימות גילוי המודלים של Claude Code יכללו מודלים שאינם של Claude. אזהרה: מכפיל את רשומות הקטלוג עבור כל הלקוחות כאשר זה מופעל באופן גלובלי.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "הפעל נתיבי קבלה וירטואליים אדפטיביים לכל דייר (tenant) עבור שליחת ספקים (#9654): פרץ עומס של דייר אחד כבר לא מחזיר 503 לדייר אחר. משתנה הסביבה OMNIROUTE_CHAT_VIRTUAL_LANES גובר על הגדרה זו בלוח הבקרה; השינויים נכנסים לתוקף לאחר הפעלת השרת מחדש.", diff --git a/src/i18n/messages/hi.json b/src/i18n/messages/hi.json index f0343fbc6c3d..a59140771989 100644 --- a/src/i18n/messages/hi.json +++ b/src/i18n/messages/hi.json @@ -987,6 +987,7 @@ "disabled": "अक्षम", "featureFlagOmnirouteEmergencyFallbackDescription": "बजट समाप्त हो चुके अनुरोधों को आपातकालीन निःशुल्क फ़ॉलबैक प्रदाता/मॉडल पर रूट करें।", "featureFlagArenaEloSyncEnabledDescription": "मॉडल इंटेलिजेंस रैंकिंग के लिए आवधिक Arena AI लीडरबोर्ड ELO सिंक सक्षम करें।", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models पर claude/<provider>/<model> मिरर आईडी का विज्ञापन करें ताकि Claude Code गेटवे मॉडल खोज सूची में गैर-Claude मॉडल शामिल हो सकें। चेतावनी: जब वैश्विक रूप से सक्षम किया जाता है तो सभी ग्राहकों के लिए कैटलॉग प्रविष्टियों को डबल करता है।", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "प्रदाता डिस्पैच के लिए प्रति-टेनेंट अनुकूली वर्चुअल एडमिशन लेन सक्षम करें (#9654): एक टेनेंट का बर्स्ट अब दूसरे टेनेंट को 503 नहीं देता। OMNIROUTE_CHAT_VIRTUAL_LANES पर्यावरण चर इस डैशबोर्ड सेटिंग पर प्राथमिकता रखता है; परिवर्तन सर्वर पुनः आरंभ पर प्रभावी होते हैं।", diff --git a/src/i18n/messages/hu.json b/src/i18n/messages/hu.json index c3debac89bf9..67e5ef7e0cd4 100644 --- a/src/i18n/messages/hu.json +++ b/src/i18n/messages/hu.json @@ -987,6 +987,7 @@ "disabled": "Letiltva", "featureFlagOmnirouteEmergencyFallbackDescription": "A keretet kimerítő kérések átirányítása a vészhelyzeti ingyenes tartalék szolgáltatóhoz/modellhez.", "featureFlagArenaEloSyncEnabledDescription": "Rendszeres Arena AI ranglista ELO szinkronizáció engedélyezése a modellintelligencia rangsorokhoz.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Hirdesse a claude/<provider>/<model> tükör azonosítókat a /v1/models-on, hogy a Claude Code átjáró modell felfedezése nem Claude modelleket is listázzon. Figyelmeztetés: globális engedélyezés esetén megduplázza a katalógus bejegyzéseket minden kliens számára.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Tegye lehetővé a bérlőnkénti adaptív virtuális beléptetősávokat a szolgáltatók felé történő továbbításhoz (#9654): az egyik bérlő kiugró terhelése már nem okoz 503-as hibát egy másiknál. Az OMNIROUTE_CHAT_VIRTUAL_LANES környezeti változó felülírja ezt a vezérlőpult-beállítást; a változtatások a szerver újraindításakor lépnek életbe.", diff --git a/src/i18n/messages/id.json b/src/i18n/messages/id.json index dc84d67f5913..f6914ded4d6c 100644 --- a/src/i18n/messages/id.json +++ b/src/i18n/messages/id.json @@ -987,6 +987,7 @@ "disabled": "Dinonaktifkan", "featureFlagOmnirouteEmergencyFallbackDescription": "Arahkan permintaan yang kehabisan anggaran ke penyedia/model fallback gratis darurat.", "featureFlagArenaEloSyncEnabledDescription": "Aktifkan sinkronisasi ELO papan peringkat Arena AI berkala untuk peringkat kecerdasan model.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Iklankan claude/<provider>/<model> mirror ids di /v1/models sehingga daftar penemuan model gateway Claude Code mencantumkan model non-Claude. Peringatan: menggandakan entri katalog untuk semua klien saat diaktifkan secara global.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktifkan jalur penerimaan virtual adaptif per-tenant untuk pengiriman penyedia (#9654): lonjakan satu tenant tidak lagi mengembalikan 503 ke tenant lain. Variabel lingkungan OMNIROUTE_CHAT_VIRTUAL_LANES menang atas pengaturan dasbor ini; perubahan berlaku setelah server dimulai ulang.", diff --git a/src/i18n/messages/in.json b/src/i18n/messages/in.json index 448791f54b51..4029832c7628 100644 --- a/src/i18n/messages/in.json +++ b/src/i18n/messages/in.json @@ -987,6 +987,7 @@ "disabled": "Dinonaktifkan", "featureFlagOmnirouteEmergencyFallbackDescription": "Rute permintaan yang kehabisan anggaran ke penyedia/model cadangan gratis darurat.", "featureFlagArenaEloSyncEnabledDescription": "Aktifkan sinkronisasi ELO papan peringkat Arena AI berkala untuk peringkat kecerdasan model.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Iklankan claude/<provider>/<model> mirror ids di /v1/models sehingga daftar penemuan model gateway Claude Code mencantumkan model non-Claude. Peringatan: menggandakan entri katalog untuk semua klien saat diaktifkan secara global.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktifkan jalur penerimaan virtual adaptif per-tenant untuk pengiriman penyedia (#9654): lonjakan satu tenant tidak lagi mengembalikan 503 ke tenant lain. Variabel lingkungan OMNIROUTE_CHAT_VIRTUAL_LANES menang atas pengaturan dasbor ini; perubahan berlaku setelah server dimulai ulang.", diff --git a/src/i18n/messages/it.json b/src/i18n/messages/it.json index a23142656464..a85c6de5933e 100644 --- a/src/i18n/messages/it.json +++ b/src/i18n/messages/it.json @@ -987,6 +987,7 @@ "disabled": "Disabilitato", "featureFlagOmnirouteEmergencyFallbackDescription": "Indirizza le richieste con budget esaurito al provider/modello di fallback gratuito di emergenza.", "featureFlagArenaEloSyncEnabledDescription": "Abilita la sincronizzazione periodica dell'ELO della classifica Arena AI per le graduatorie di intelligenza dei modelli.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Mostra gli id specchio claude/<provider>/<model> su /v1/models in modo che la scoperta dei modelli gateway di Claude Code elenchi i modelli non-Claude. Attenzione: raddoppia le voci nel catalogo per tutti i client quando abilitato globalmente.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Attiva corsie di ammissione virtuali adattive per tenant per l'invio ai provider (#9654): il picco di un tenant non restituisce più 503 a un altro. La variabile d'ambiente OMNIROUTE_CHAT_VIRTUAL_LANES ha la precedenza su questa impostazione della dashboard; le modifiche hanno effetto al riavvio del server.", diff --git a/src/i18n/messages/ja.json b/src/i18n/messages/ja.json index 6670b9c7f276..80e40c409945 100644 --- a/src/i18n/messages/ja.json +++ b/src/i18n/messages/ja.json @@ -987,6 +987,7 @@ "disabled": "無効", "featureFlagOmnirouteEmergencyFallbackDescription": "予算を使い果たしたリクエストを、緊急用の無料フォールバックプロバイダー/モデルにルーティングします。", "featureFlagArenaEloSyncEnabledDescription": "モデルのインテリジェンスランキング向けに、定期的な Arena AI リーダーボード ELO 同期を有効にします。", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models で Claude Code ゲートウェイのモデル発見リストに非 Claude モデルを表示するために、claude/<provider>/<model> ミラー ID を広告します。警告: グローバルに有効にすると、すべてのクライアントのカタログエントリが重複します。", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "プロバイダーへのディスパッチ用に、テナントごとの適応型仮想受付レーンを有効にします（#9654）：あるテナントのバーストが他のテナントに503を返さなくなります。OMNIROUTE_CHAT_VIRTUAL_LANES環境変数はこのダッシュボード設定より優先されます。変更はサーバー再起動時に反映されます。", diff --git a/src/i18n/messages/ko.json b/src/i18n/messages/ko.json index 0b42613cc320..c7bd51cb9d05 100644 --- a/src/i18n/messages/ko.json +++ b/src/i18n/messages/ko.json @@ -987,6 +987,7 @@ "disabled": "비활성화됨", "featureFlagOmnirouteEmergencyFallbackDescription": "예산이 소진된 요청을 긴급 무료 폴백 제공자/모델로 라우팅합니다.", "featureFlagArenaEloSyncEnabledDescription": "모델 지능 순위를 위해 주기적인 Arena AI 리더보드 ELO 동기화를 활성화합니다.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models에서 Claude Code 게이트웨이 모델 검색 목록에 비Claude 모델이 포함되도록 claude/<provider>/<model> 미러 ID를 광고합니다. 경고: 전역적으로 활성화하면 모든 클라이언트에 대해 카탈로그 항목이 두 배로 증가합니다.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "공급자 디스패치를 위해 테넌트별 적응형 가상 승인 레인을 활성화합니다(#9654): 한 테넌트의 폭증이 더 이상 다른 테넌트에 503을 반환하지 않습니다. OMNIROUTE_CHAT_VIRTUAL_LANES 환경 변수가 이 대시보드 설정보다 우선하며, 변경 사항은 서버 재시작 시 적용됩니다.", diff --git a/src/i18n/messages/mr.json b/src/i18n/messages/mr.json index 45ebea91d4f4..5a0a628940dd 100644 --- a/src/i18n/messages/mr.json +++ b/src/i18n/messages/mr.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models वर claude/<provider>/<model> मिरर आयडीज जाहिरात करा जेणेकरून Claude Code गेटवे मॉडेल शोध सूचीमध्ये नॉन-Claude मॉडेल्स समाविष्ट होतील. चेतावणी: जागतिक स्तरावर सक्षम केल्यास सर्व क्लायंटसाठी कॅटलॉग नोंदी दुहेरी होतात.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "प्रदाता डिस्पॅचसाठी प्रति-टेनंट अनुकूली व्हर्च्युअल अॅडमिशन लेन सक्षम करा (#9654): एका टेनंटचा बर्स्ट यापुढे दुसऱ्या टेनंटला 503 देत नाही. OMNIROUTE_CHAT_VIRTUAL_LANES पर्यावरण चल या डॅशबोर्ड सेटिंगपेक्षा वरचढ आहे; बदल सर्व्हर रीस्टार्ट केल्यावर प्रभावी होतात.", diff --git a/src/i18n/messages/ms.json b/src/i18n/messages/ms.json index cd4b9060bba5..5349f3fc08c9 100644 --- a/src/i18n/messages/ms.json +++ b/src/i18n/messages/ms.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Iklankan claude/<provider>/<model> mirror ids pada /v1/models supaya senarai penemuan model gerbang Claude Code termasuk model bukan Claude. Amaran: menggandakan entri katalog untuk semua klien apabila diaktifkan secara global.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktifkan lorong kemasukan maya adaptif setiap-tenant untuk penghantaran pembekal (#9654): lonjakan satu tenant tidak lagi memberikan 503 kepada tenant lain. Pemboleh ubah persekitaran OMNIROUTE_CHAT_VIRTUAL_LANES mengatasi tetapan papan pemuka ini; perubahan berkuat kuasa apabila pelayan dimulakan semula.", diff --git a/src/i18n/messages/nl.json b/src/i18n/messages/nl.json index 90344e67c24c..676f55885f78 100644 --- a/src/i18n/messages/nl.json +++ b/src/i18n/messages/nl.json @@ -987,6 +987,7 @@ "disabled": "Uitgeschakeld", "featureFlagOmnirouteEmergencyFallbackDescription": "Routeer verzoeken met uitgeput budget naar de gratis nood-fallbackprovider/-model.", "featureFlagArenaEloSyncEnabledDescription": "Schakel periodieke ELO-synchronisatie van het Arena AI-leaderboard in voor modelintelligentieranglijsten.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Adverteer claude/<provider>/<model> spiegel-id's op /v1/models zodat Claude Code gateway modelontdekking niet-Claude modellen vermeldt. Waarschuwing: dubbele catalogusvermeldingen voor alle klanten wanneer wereldwijd ingeschakeld.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Schakel adaptieve virtuele toegangsbanen per tenant in voor provider-dispatch (#9654): een piek van de ene tenant geeft de andere niet langer een 503. De omgevingsvariabele OMNIROUTE_CHAT_VIRTUAL_LANES wint het van deze dashboard-instelling; wijzigingen gaan in bij een serverherstart.", diff --git a/src/i18n/messages/no.json b/src/i18n/messages/no.json index bf7812a39327..dd8e47b3bb0b 100644 --- a/src/i18n/messages/no.json +++ b/src/i18n/messages/no.json @@ -987,6 +987,7 @@ "disabled": "Deaktivert", "featureFlagOmnirouteEmergencyFallbackDescription": "Rut forespørsler med oppbrukt budsjett til gratis reserveleverandør/-modell for nødstilfeller.", "featureFlagArenaEloSyncEnabledDescription": "Aktiver periodisk synkronisering av Arena AI-ledertavlens ELO for rangering av modellintelligens.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Reklamer claude/<provider>/<model> speil-id-er på /v1/models slik at Claude Code gateway-modelloppdagelse viser ikke-Claude-modeller. Advarsel: dobler katalogoppføringer for alle klienter når det er aktivert globalt.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktiver adaptive virtuelle tilgangsfelt per tenant for leverandørdistribusjon (#9654): et utbrudd fra én tenant gir ikke lenger en annen 503. Miljøvariabelen OMNIROUTE_CHAT_VIRTUAL_LANES overstyrer denne innstillingen i dashbordet; endringer trer i kraft ved omstart av serveren.", diff --git a/src/i18n/messages/phi.json b/src/i18n/messages/phi.json index a5b7af7efed6..085c48ca5654 100644 --- a/src/i18n/messages/phi.json +++ b/src/i18n/messages/phi.json @@ -987,6 +987,7 @@ "disabled": "Naka-disable", "featureFlagOmnirouteEmergencyFallbackDescription": "I-route ang mga request na naubusan ng budget sa emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "I-enable ang pana-panahong Arena AI leaderboard ELO sync para sa mga ranking ng intelligence ng modelo.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "I-anunsyo ang claude/<provider>/<model> mirror ids sa /v1/models upang ang Claude Code gateway model discovery ay maglista ng mga non-Claude models. Babala: nagdodoble ng catalog entries para sa lahat ng kliyente kapag pinagana nang globally.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Paganahin ang adaptive virtual admission lanes para sa bawat tenant sa pagpapadala ng provider (#9654): ang pag-akyat ng trapiko ng isang tenant ay hindi na nagbibigay ng 503 sa iba. Ang environment variable na OMNIROUTE_CHAT_VIRTUAL_LANES ay mas nangingibabaw sa setting na ito sa dashboard; magkakabisa ang mga pagbabago sa pag-restart ng server.", diff --git a/src/i18n/messages/pl.json b/src/i18n/messages/pl.json index d7188622187b..5c3f239ec166 100644 --- a/src/i18n/messages/pl.json +++ b/src/i18n/messages/pl.json @@ -987,6 +987,7 @@ "disabled": "Wyłączone", "featureFlagOmnirouteEmergencyFallbackDescription": "Kierowanie żądań z wyczerpanym budżetem do awaryjnego, bezpłatnego fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Włączenie okresowej synchronizacji ELO tabeli liderów Arena AI dla rankingów inteligencji model.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Reklamuj identyfikatory luster claude/<provider>/<model> na /v1/models, aby brama modelu Claude Code wyświetlała listę modeli niebędących Claude. Uwaga: podwaja wpisy w katalogu dla wszystkich klientów, gdy jest włączone globalnie.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Włącz adaptacyjne wirtualne pasma przyjęć dla każdego tenanta przy wysyłce do dostawców (#9654): przeciążenie jednego tenanta nie powoduje już błędu 503 u innego. Zmienna środowiskowa OMNIROUTE_CHAT_VIRTUAL_LANES ma pierwszeństwo przed tym ustawieniem w panelu; zmiany wchodzą w życie po restarcie serwera.", diff --git a/src/i18n/messages/pt-BR.json b/src/i18n/messages/pt-BR.json index 7b770314c112..6b6a669b2349 100644 --- a/src/i18n/messages/pt-BR.json +++ b/src/i18n/messages/pt-BR.json @@ -988,6 +988,7 @@ "featureFlagOmnirouteEmergencyFallbackDescription": "Roteie solicitações com orçamento esgotado para o provedor/modelo de fallback gratuito de emergência.", "featureFlagOmnirouteDisableThinkingLevelVariantsDescription": "Desative a geração de variantes de nível de pensamento (por exemplo, -low, -medium, -high) no catálogo /v1/models.", "featureFlagArenaEloSyncEnabledDescription": "Ativar a sincronização periódica do ELO do leaderboard da Arena AI para classificações de inteligência do modelo.", + "featureFlagUniversalContextHandoffEnabledDescription": "Gera e injeta resumos da conversa quando o roteamento de combo troca de modelo. Desative para tratar trocas de modelo de forma independente e evitar requisições de handoff em segundo plano para todos os combos existentes e futuros.", "featureFlagExposeCcDiscoveryAliasesDescription": "Divulgar ids espelho claude/<provider>/<model> em /v1/models para que a descoberta de modelos do gateway Claude Code liste modelos não-Claude. Atenção: duplica as entradas do catálogo para todos os clientes quando ativado globalmente.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Ative faixas de admissão virtuais adaptativas por tenant para o despacho de provedores (#9654): o pico de um tenant não gera mais 503 para outro. A variável de ambiente OMNIROUTE_CHAT_VIRTUAL_LANES tem precedência sobre esta configuração do painel; as alterações entram em vigor ao reiniciar o servidor.", diff --git a/src/i18n/messages/pt.json b/src/i18n/messages/pt.json index 249d3c13b074..51ca10e9cdc4 100644 --- a/src/i18n/messages/pt.json +++ b/src/i18n/messages/pt.json @@ -987,6 +987,7 @@ "disabled": "Desativado", "featureFlagOmnirouteEmergencyFallbackDescription": "Encaminhar pedidos com orçamento esgotado para o fornecedor/modelo de contingência gratuito de emergência.", "featureFlagArenaEloSyncEnabledDescription": "Ativar a sincronização periódica do ELO da tabela de classificação da Arena AI para classificações de inteligência do modelo.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Anuncie os ids de espelho claude/<provider>/<model> em /v1/models para que a descoberta de modelos do gateway Claude Code liste modelos não Claude. Aviso: duplica entradas de catálogo para todos os clientes quando ativado globalmente.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Ative filas de admissão virtuais adaptativas por tenant para o encaminhamento de fornecedores (#9654): um pico de tráfego de um tenant já não gera 503 noutro. A variável de ambiente OMNIROUTE_CHAT_VIRTUAL_LANES sobrepõe-se a esta definição do painel; as alterações entram em vigor ao reiniciar o servidor.", diff --git a/src/i18n/messages/ro.json b/src/i18n/messages/ro.json index e110eef0c753..3e0bb710f336 100644 --- a/src/i18n/messages/ro.json +++ b/src/i18n/messages/ro.json @@ -987,6 +987,7 @@ "disabled": "Dezactivat", "featureFlagOmnirouteEmergencyFallbackDescription": "Redirecționează cererile cu buget epuizat către furnizorul/modelul de rezervă gratuit de urgență.", "featureFlagArenaEloSyncEnabledDescription": "Activează sincronizarea periodică ELO a clasamentului Arena AI pentru clasamentele de inteligență ale modelelor.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Publica id-urile mirror claude/<provider>/<model> pe /v1/models astfel încât lista de descoperire a modelului Claude Code să includă modele non-Claude. Atenție: dublează intrările din catalog pentru toți clienții când este activat global.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Activați benzile de admitere virtuale adaptive per-tenant pentru expedierea către furnizori (#9654): un vârf de trafic al unui tenant nu mai returnează 503 altui tenant. Variabila de mediu OMNIROUTE_CHAT_VIRTUAL_LANES are prioritate față de această setare din panou; modificările intră în vigoare la repornirea serverului.", diff --git a/src/i18n/messages/ru.json b/src/i18n/messages/ru.json index 32ca8427755f..45b244538646 100644 --- a/src/i18n/messages/ru.json +++ b/src/i18n/messages/ru.json @@ -987,6 +987,7 @@ "disabled": "Отключено", "featureFlagOmnirouteEmergencyFallbackDescription": "Перенаправлять запросы при исчерпании бюджета на резервный бесплатный провайдер/модель.", "featureFlagArenaEloSyncEnabledDescription": "Включить периодическую синхронизацию ELO из таблицы лидеров Arena AI для рейтинга интеллектуальности моделей.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Публиковать claude/<провайдер>/<модель> зеркальные ID на /v1/models, чтобы Claude Code мог видеть не-Claude модели. Внимание: удваивает записи каталога для всех клиентов при глобальном включении.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Включите адаптивные виртуальные полосы допуска для каждого тенанта при маршрутизации к провайдерам (#9654): всплеск нагрузки одного тенанта больше не вызывает 503 у другого. Переменная окружения OMNIROUTE_CHAT_VIRTUAL_LANES имеет приоритет над этой настройкой в панели; изменения вступают в силу после перезапуска сервера.", diff --git a/src/i18n/messages/sk.json b/src/i18n/messages/sk.json index a97a64ae240a..a4d14e20b013 100644 --- a/src/i18n/messages/sk.json +++ b/src/i18n/messages/sk.json @@ -987,6 +987,7 @@ "disabled": "Zakázané", "featureFlagOmnirouteEmergencyFallbackDescription": "Smerovať požiadavky s vyčerpaným rozpočtom na núdzového bezplatného záložného poskytovateľa/model.", "featureFlagArenaEloSyncEnabledDescription": "Povoliť pravidelnú synchronizáciu ELO z rebríčka Arena AI pre hodnotenie inteligencie modelov.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Inzerujte claude/<provider>/<model> zrkadlové ID na /v1/models, aby zoznam objavovania modelov Claude Code obsahoval aj modely, ktoré nie sú Claude. Upozornenie: pri globálnom povolení zdvojuje záznamy v katalógu pre všetkých klientov.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Povoľte adaptívne virtuálne vstupné pruhy pre každého nájomcu (tenant) pri odosielaní poskytovateľom (#9654): špička jedného nájomcu už nespôsobí 503 u iného. Premenná prostredia OMNIROUTE_CHAT_VIRTUAL_LANES má prednosť pred týmto nastavením v riadiacom paneli; zmeny sa prejavia po reštarte servera.", diff --git a/src/i18n/messages/sv.json b/src/i18n/messages/sv.json index 29488723d837..acac6435d499 100644 --- a/src/i18n/messages/sv.json +++ b/src/i18n/messages/sv.json @@ -987,6 +987,7 @@ "disabled": "Inaktiverad", "featureFlagOmnirouteEmergencyFallbackDescription": "Dirigera anrop med förbrukad budget till den kostnadsfria reservleverantören/-modellen för nödfall.", "featureFlagArenaEloSyncEnabledDescription": "Aktivera periodisk synkronisering av Arena AI-topplistans ELO för rankning av modellintelligens.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Reklamera claude/<provider>/<model> spegel-id på /v1/models så att Claude Code gateway-modellens upptäcktslista visar icke-Claude-modeller. Varning: dubblerar katalogposter för alla klienter när det är aktiverat globalt.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Aktivera adaptiva virtuella åtkomstfiler per tenant för providerutskick (#9654): en tenants burst ger inte längre en annan 503. Miljövariabeln OMNIROUTE_CHAT_VIRTUAL_LANES har företräde framför den här inställningen i instrumentpanelen; ändringarna träder i kraft vid omstart av servern.", diff --git a/src/i18n/messages/sw.json b/src/i18n/messages/sw.json index fc28cefedbd6..bb348a22cb17 100644 --- a/src/i18n/messages/sw.json +++ b/src/i18n/messages/sw.json @@ -987,6 +987,7 @@ "disabled": "Imezimwa", "featureFlagOmnirouteEmergencyFallbackDescription": "Elekeza maombi yaliyomaliza bajeti kwenye mtoa huduma/muundo wa dharura wa akiba usiolipiwa.", "featureFlagArenaEloSyncEnabledDescription": "Washa usawazishaji wa mara kwa mara wa ELO wa ubao wa wanaoongoza wa Arena AI kwa viwango vya akili vya muundo.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Tangaza claude/<provider>/<model> vitambulisho vya kioo kwenye /v1/models ili orodha ya kugundua modeli za Claude Code iwe na modeli zisizo za Claude. Onyo: inafanya kuingia mara mbili kwenye katalogi kwa wateja wote inapowekwa kuwa ya ulimwengu mzima.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Washa njia za uandikishaji pepe zinazobadilika kwa kila mpangaji (tenant) kwa utumaji wa watoa huduma (#9654): mlipuko wa mpangaji mmoja hautoi tena 503 kwa mwingine. Kigezo cha mazingira cha OMNIROUTE_CHAT_VIRTUAL_LANES kinashinda mpangilio huu wa dashibodi; mabadiliko yanatumika wakati seva inapoanzishwa upya.", diff --git a/src/i18n/messages/ta.json b/src/i18n/messages/ta.json index e4343dcffbbb..d94d9ffbc211 100644 --- a/src/i18n/messages/ta.json +++ b/src/i18n/messages/ta.json @@ -987,6 +987,7 @@ "disabled": "முடக்கப்பட்டது", "featureFlagOmnirouteEmergencyFallbackDescription": "பட்ஜெட் தீர்ந்த கோரிக்கைகளை அவசரகால இலவச ஃபால்பேக் வழங்குநர்/மாடலுக்கு வழிசெலுத்துங்கள்.", "featureFlagArenaEloSyncEnabledDescription": "மாடல் நுண்ணறிவு தரவரிசைகளுக்காக அவ்வப்போதான Arena AI லீடர்போர்டு ELO ஒத்திசைவை இயக்குங்கள்.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models இல் Claude Code gateway மாதிரி கண்டுபிடிப்பு பட்டியலில் non-Claude மாதிரிகளை காட்ட Claude/<provider>/<model> மின்னூல் அடையாளங்களை விளம்பரம் செய்யவும். எச்சரிக்கை: உலகளாவியமாக செயல்படுத்தப்பட்டால் அனைத்து கிளையன்டுகளுக்கும் பட்டியல் பதிவுகளை இரட்டைப்படுத்துகிறது.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "வழங்குநர் அனுப்பீட்டிற்கு ஒவ்வொரு குத்தகைதாரருக்கும் (tenant) தகவமைப்பு மெய்நிகர் சேர்க்கைப் பாதைகளை இயக்கு (#9654): ஒரு குத்தகைதாரரின் அதிகரிப்பு இனி மற்றொருவருக்கு 503 ஐ அளிக்காது. OMNIROUTE_CHAT_VIRTUAL_LANES சூழல் மாறி இந்த டாஷ்போர்டு அமைப்பை விட முன்னுரிமை பெறுகிறது; மாற்றங்கள் சேவையகம் மறுதொடக்கத்தில் நடைமுறைக்கு வரும்.", diff --git a/src/i18n/messages/te.json b/src/i18n/messages/te.json index 510db6a4f7b5..d904c4911e5e 100644 --- a/src/i18n/messages/te.json +++ b/src/i18n/messages/te.json @@ -987,6 +987,7 @@ "disabled": "నిలిపివేయబడింది", "featureFlagOmnirouteEmergencyFallbackDescription": "బడ్జెట్ ముగిసిపోయిన అభ్యర్థనలను అత్యవసర ఉచిత ఫాల్‌బ్యాక్ ప్రొవైడర్/మోడల్‌కు రూట్ చేయండి.", "featureFlagArenaEloSyncEnabledDescription": "మోడల్ ఇంటెలిజెన్స్ ర్యాంకింగ్‌ల కోసం క్రమానుగత Arena AI లీడర్‌బోర్డ్ ELO సమకాలీకరణను ప్రారంభించండి.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models లో claude/<provider>/<model> మిర్రర్ ఐడీలను ప్రచారం చేయండి కాబట్టి Claude Code గేట్వే మోడల్ డిస్కవరీ non-Claude మోడళ్లను జాబితా చేస్తుంది. హెచ్చరిక: ఇది ప్రపంచవ్యాప్తంగా ప్రారంభించినప్పుడు అన్ని క్లయింట్ల కోసం కాటలాగ్ ఎంట్రీలను డబుల్ చేస్తుంది.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "ప్రొవైడర్ డిస్పాచ్ కోసం ప్రతి-టెనెంట్ అడాప్టివ్ వర్చువల్ అడ్మిషన్ లేన్లను ప్రారంభించండి (#9654): ఒక టెనెంట్ బర్స్ట్ ఇకపై మరొక టెనెంట్కు 503 ఇవ్వదు. OMNIROUTE_CHAT_VIRTUAL_LANES ఎన్విరాన్మెంట్ వేరియబుల్ ఈ డాష్బోర్డ్ సెట్టింగ్ కంటే ప్రాధాన్యత పొందుతుంది; మార్పులు సర్వర్ పునఃప్రారంభంలో ప్రభావం చూపుతాయి.", diff --git a/src/i18n/messages/th.json b/src/i18n/messages/th.json index 26b03358314c..bf27c1a69c9d 100644 --- a/src/i18n/messages/th.json +++ b/src/i18n/messages/th.json @@ -987,6 +987,7 @@ "disabled": "ปิดใช้งาน", "featureFlagOmnirouteEmergencyFallbackDescription": "กำหนดเส้นทางคำขอที่งบประมาณหมดไปยังผู้ให้บริการ/โมเดลสำรองฟรีในกรณีฉุกเฉิน", "featureFlagArenaEloSyncEnabledDescription": "เปิดใช้งานการซิงค์ ELO ของลีดเดอร์บอร์ด Arena AI เป็นระยะสำหรับการจัดอันดับความฉลาดของโมเดล", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "โฆษณา claude/<provider>/<model> mirror ids บน /v1/models เพื่อให้รายการการค้นหาโมเดลของ Claude Code แสดงโมเดลที่ไม่ใช่ Claude เตือน: จะทำให้มีรายการในแคตตาล็อกซ้ำสำหรับลูกค้าทุกคนเมื่อเปิดใช้งานทั่วโลก.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "เปิดใช้เลนรับเข้าเสมือนแบบปรับตัวต่อเทนแนนต์สำหรับการส่งไปยังผู้ให้บริการ (#9654): การพุ่งสูงของเทนแนนต์หนึ่งจะไม่ทำให้อีกเทนแนนต์ได้รับ 503 อีกต่อไป ตัวแปรสภาพแวดล้อม OMNIROUTE_CHAT_VIRTUAL_LANES มีผลเหนือการตั้งค่าแดชบอร์ดนี้ การเปลี่ยนแปลงมีผลเมื่อรีสตาร์ทเซิร์ฟเวอร์", diff --git a/src/i18n/messages/tr.json b/src/i18n/messages/tr.json index 59c4a43ca47b..b4a7b552dcba 100644 --- a/src/i18n/messages/tr.json +++ b/src/i18n/messages/tr.json @@ -987,6 +987,7 @@ "disabled": "Devre dışı", "featureFlagOmnirouteEmergencyFallbackDescription": "Bütçesi tükenmiş istekleri acil durum ücretsiz yedek sağlayıcıya/modele yönlendirin.", "featureFlagArenaEloSyncEnabledDescription": "Model zekası sıralamaları için periyodik Arena AI liderlik tablosu ELO senkronizasyonunu etkinleştirin.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models üzerinde claude/<provider>/<model> ayna kimliklerini tanıtın, böylece Claude Code geçidi model keşfi, Claude olmayan modelleri listeler. Uyarı: Küresel olarak etkinleştirildiğinde tüm istemciler için katalog girişlerini iki katına çıkarır.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Sağlayıcı gönderimi için kiracı başına uyarlanabilir sanal kabul şeritlerini etkinleştirin (#9654): bir kiracının ani yükü artık diğerinde 503 hatasına neden olmaz. OMNIROUTE_CHAT_VIRTUAL_LANES ortam değişkeni bu panel ayarına göre önceliklidir; değişiklikler sunucu yeniden başlatıldığında geçerli olur.", diff --git a/src/i18n/messages/uk-UA.json b/src/i18n/messages/uk-UA.json index eb6ef025c4d8..73e23623d0e0 100644 --- a/src/i18n/messages/uk-UA.json +++ b/src/i18n/messages/uk-UA.json @@ -987,6 +987,7 @@ "disabled": "Вимкнено", "featureFlagOmnirouteEmergencyFallbackDescription": "Перенаправляти запити з вичерпаним бюджетом на резервного безкоштовного провайдера/модель.", "featureFlagArenaEloSyncEnabledDescription": "Увімкнути періодичну синхронізацію ELO з таблиці лідерів Arena AI для рейтингу інтелекту моделей.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "Рекламуйте claude/<provider>/<model> mirror ids на /v1/models, щоб модель виявлення Claude Code gateway перераховувала не Claude моделі. Увага: подвоює записи каталогу для всіх клієнтів, коли увімкнено глобально.", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "Увімкніть адаптивні віртуальні смуги допуску для кожного тенанта під час надсилання провайдерам (#9654): сплеск навантаження одного тенанта більше не викликає 503 в іншого. Змінна середовища OMNIROUTE_CHAT_VIRTUAL_LANES має пріоритет над цим налаштуванням у панелі; зміни набувають чинності після перезапуску сервера.", diff --git a/src/i18n/messages/ur.json b/src/i18n/messages/ur.json index ecac5436307f..0fbf216cbd70 100644 --- a/src/i18n/messages/ur.json +++ b/src/i18n/messages/ur.json @@ -987,6 +987,7 @@ "disabled": "Disabled", "featureFlagOmnirouteEmergencyFallbackDescription": "Route budget-exhausted requests to the emergency free fallback provider/model.", "featureFlagArenaEloSyncEnabledDescription": "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "/v1/models پر claude/<provider>/<model> آئینہ شناختوں کا اشتہار دیں تاکہ Claude Code گیٹ وے ماڈل کی دریافت غیر-Claude ماڈلز کی فہرست بنائے۔ انتباہ: جب عالمی طور پر فعال ہو تو تمام کلائنٹس کے لیے کیٹلاگ کی اندراجات دوگنا کرتا ہے۔", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "پرووائیڈر بھیجنے کے لیے فی ٹیننٹ انکولی ورچوئل ایڈمیشن لین فعال کریں (#9654): ایک ٹیننٹ کا اچانک بوجھ اب دوسرے ٹیننٹ کو 503 نہیں دیتا۔ OMNIROUTE_CHAT_VIRTUAL_LANES ماحولیاتی متغیر اس ڈیش بورڈ سیٹنگ پر فوقیت رکھتا ہے؛ تبدیلیاں سرور دوبارہ شروع ہونے پر اثر انداز ہوتی ہیں۔", diff --git a/src/i18n/messages/vi.json b/src/i18n/messages/vi.json index bc62bf2f234f..e3610dfa0f9f 100644 --- a/src/i18n/messages/vi.json +++ b/src/i18n/messages/vi.json @@ -988,6 +988,7 @@ "featureFlagOmnirouteEmergencyFallbackDescription": "Định tuyến các yêu cầu đã hết ngân sách đến nhà cung cấp/mô hình dự phòng khẩn cấp miễn phí.", "featureFlagOmnirouteDisableThinkingLevelVariantsDescription": "Tắt tính năng tạo biến thể cấp độ suy nghĩ (ví dụ: -low, -medium, -high) trong danh mục /v1/models.", "featureFlagArenaEloSyncEnabledDescription": "Bật đồng bộ ELO định kỳ từ bảng xếp hạng Arena AI để xếp hạng năng lực của mô hình.", + "featureFlagUniversalContextHandoffEnabledDescription": "Tạo và chèn bản tóm tắt hội thoại khi định tuyến combo chuyển đổi mô hình. Tắt để xử lý các lần chuyển mô hình một cách độc lập và ngăn các yêu cầu bàn giao chạy nền cho mọi combo hiện có và trong tương lai.", "featureFlagExposeCcDiscoveryAliasesDescription": "Quảng bá các id phản chiếu claude/<provider>/<model> trên /v1/models để tính năng khám phá mô hình qua gateway của Claude Code liệt kê được các mô hình không phải Claude. Cảnh báo: khi bật ở phạm vi toàn cục, số mục trong danh mục tăng gấp đôi với mọi client.", "featureFlagNoThinkingAliasEnabledDescription": "Công tắc chính cho các bí danh gateway no-think/<provider>/<model>. Bật (mặc định): /v1/models quảng bá biến thể không suy nghĩ cho mọi mô hình Claude có khả năng suy nghĩ đủ điều kiện, và id no-think/ được gửi trên một yêu cầu sẽ giải quyết lại về mô hình thực với phần lý luận bị triệt tiêu. Tắt: không có biến thể nào được quảng bá và id no-think/ được xử lý như bất kỳ id mô hình không xác định nào khác. Tùy chọn tham gia/từ chối ModelSpec.noThinkingAlias theo từng mô hình vẫn áp dụng khi tính năng này bật.", "featureFlagChatVirtualLanesEnabledDescription": "Bật làn tiếp nhận ảo thích ứng cho từng đối tượng thuê (tenant) để phân phối nhà cung cấp (#9654): một đợt bùng phát của tenant này không còn trả 503 cho tenant khác. Biến môi trường OMNIROUTE_CHAT_VIRTUAL_LANES được ưu tiên hơn cài đặt bảng điều khiển này; các thay đổi có hiệu lực khi khởi động lại máy chủ.", diff --git a/src/i18n/messages/zh-CN.json b/src/i18n/messages/zh-CN.json index 503597b363eb..ceed86d11e4e 100644 --- a/src/i18n/messages/zh-CN.json +++ b/src/i18n/messages/zh-CN.json @@ -987,6 +987,7 @@ "disabled": "已禁用", "featureFlagOmnirouteEmergencyFallbackDescription": "将预算耗尽的请求路由到紧急免费备用提供者/模型。", "featureFlagArenaEloSyncEnabledDescription": "启用定期同步 Arena AI 排行榜 ELO，用于模型智能排名。", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "在 /v1/models 上发布 claude/<provider>/<model> 镜像 ID，让 Claude Code 网关模型发现能列出非 Claude 模型。警告：全局启用会使所有客户端的目录条目翻倍。", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "为提供者调度启用按租户的自适应虚拟准入通道（#9654）：一个租户的突发流量不再导致另一个租户收到 503。OMNIROUTE_CHAT_VIRTUAL_LANES 环境变量优先于此仪表板设置；更改在服务器重启后生效。", diff --git a/src/i18n/messages/zh-TW.json b/src/i18n/messages/zh-TW.json index b14361897299..042f8ff9646a 100644 --- a/src/i18n/messages/zh-TW.json +++ b/src/i18n/messages/zh-TW.json @@ -987,6 +987,7 @@ "disabled": "已停用", "featureFlagOmnirouteEmergencyFallbackDescription": "將預算耗盡的請求路由到緊急免費備用提供者/模型。", "featureFlagArenaEloSyncEnabledDescription": "啟用定期 Arena AI 排行榜 ELO 同步，用於模型智慧排名。", + "featureFlagUniversalContextHandoffEnabledDescription": "__MISSING__:Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", "featureFlagExposeCcDiscoveryAliasesDescription": "在 /v1/models 上廣告 claude/<provider>/<model> 鏡像 ID，以便 Claude Code 閘道模型發現列出非 Claude 模型。警告：當全域啟用時，會為所有客戶端重複目錄條目。", "featureFlagNoThinkingAliasEnabledDescription": "__MISSING__:Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.", "featureFlagChatVirtualLanesEnabledDescription": "為提供者調度啟用按租戶的自適應虛擬准入通道（#9654）：一個租戶的突發流量不再導致另一個租戶收到 503。OMNIROUTE_CHAT_VIRTUAL_LANES 環境變數優先於此儀表板設定；變更在伺服器重新啟動後生效。", diff --git a/src/shared/constants/featureFlagDefinitions.ts b/src/shared/constants/featureFlagDefinitions.ts index 07c478a5ca75..a73d4b7a254c 100644 --- a/src/shared/constants/featureFlagDefinitions.ts +++ b/src/shared/constants/featureFlagDefinitions.ts @@ -298,7 +298,19 @@ export const FEATURE_FLAG_DEFINITIONS: FeatureFlagDefinition[] = [ warningLevel: "info", }, - // ──────────────── Runtime (16) ──────────────── + // ──────────────── Runtime (17) ──────────────── + { + key: "UNIVERSAL_CONTEXT_HANDOFF_ENABLED", + label: "Universal Context Handoff", + description: + "Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.", + descriptionI18nKey: "featureFlagUniversalContextHandoffEnabledDescription", + category: "runtime", + defaultValue: "true", + type: "boolean", + requiresRestart: false, + warningLevel: "info", + }, { key: "RESPONSES_PASSTHROUGH_DROP_COMMENTARY", label: "Drop Responses Commentary", diff --git a/tests/unit/feature-flags-settings.test.ts b/tests/unit/feature-flags-settings.test.ts index bc4250c5a269..962f7ff60bfd 100644 --- a/tests/unit/feature-flags-settings.test.ts +++ b/tests/unit/feature-flags-settings.test.ts @@ -38,8 +38,8 @@ const { // no-think// gateway aliases) then bumped it from 52 to 53. // OMNIROUTE_DISABLE_THINKING_LEVEL_VARIANTS bumped it from 53 to 54; // the dead ONEPROXY_ENABLED (readerless since the 1proxy purge, #12091) -// brought it back to 53. -const EXPECTED_FEATURE_FLAG_COUNT = 53; +// brought it back to 53. UNIVERSAL_CONTEXT_HANDOFF_ENABLED bumped it to 54. +const EXPECTED_FEATURE_FLAG_COUNT = 54; // ────────────────────────────────────────────────────── // Test group 1 — Flag definitions registry diff --git a/tests/unit/universal-handoff.test.ts b/tests/unit/universal-handoff.test.ts index 8ef12d4311dc..45c06f963f65 100644 --- a/tests/unit/universal-handoff.test.ts +++ b/tests/unit/universal-handoff.test.ts @@ -20,6 +20,18 @@ test("resolveUniversalHandoffConfig returns disabled defaults when no config", ( assert.strictEqual(r.preserveSystemPrompt, true); }); +test("global feature flag can disable handoff for every combo", () => { + const previous = process.env.UNIVERSAL_CONTEXT_HANDOFF_ENABLED; + process.env.UNIVERSAL_CONTEXT_HANDOFF_ENABLED = "false"; + try { + const r = resolveUniversalHandoffConfig({ enabled: true }, { enabled: true }); + assert.strictEqual(r.enabled, false); + } finally { + if (previous === undefined) delete process.env.UNIVERSAL_CONTEXT_HANDOFF_ENABLED; + else process.env.UNIVERSAL_CONTEXT_HANDOFF_ENABLED = previous; + } +}); + test("applies combo-level config over defaults", () => { const r = resolveUniversalHandoffConfig( { enabled: true, trigger: "always", ttlMinutes: 60 } as any,