diff --git a/apps/server/src/routes/openai/v1/operations/speech-catalog/index.ts b/apps/server/src/routes/openai/v1/operations/speech-catalog/index.ts
index 6db7ccabb..d886fa117 100644
--- a/apps/server/src/routes/openai/v1/operations/speech-catalog/index.ts
+++ b/apps/server/src/routes/openai/v1/operations/speech-catalog/index.ts
@@ -125,8 +125,11 @@ export function createSpeechCatalogOperation(deps: V1RouteDeps): SpeechCatalogOp
   async function listSpeechModels() {
     // Surface the concrete TTS models the operator has configured. The UI
     // should select an explicit model id so voice catalog requests stay
-    // model-scoped instead of hiding behind DEFAULT_TTS_MODEL.
+    // model-scoped instead of hiding behind DEFAULT_TTS_MODEL. The `default`
+    // field lets the initial client selection mirror the same server-side
+    // alias that `/audio/speech` uses for `model: "auto"`.
     const config = await deps.configKV.getOrThrow('LLM_ROUTER_CONFIG')
+    const defaultModel = await deps.configKV.getOrThrow('DEFAULT_TTS_MODEL')
     // `LLM_ROUTER_CONFIG` is `optional()` at the schema, so its inferred type
     // tolerates `undefined`. `getOrThrow` already throws on missing entries,
     // so by this line we know `config` is present — the `?.` here is purely
@@ -134,6 +137,7 @@ export function createSpeechCatalogOperation(deps: V1RouteDeps): SpeechCatalogOp
     const modelIds = Object.keys(config?.tts?.models ?? {}).sort()
     return Response.json({
       models: modelIds.map(id => ({ id, name: id })),
+      default: defaultModel,
     })
   }
 
diff --git a/apps/server/src/routes/openai/v1/route.test.ts b/apps/server/src/routes/openai/v1/route.test.ts
index 4e050c234..ff115c2e3 100644
--- a/apps/server/src/routes/openai/v1/route.test.ts
+++ b/apps/server/src/routes/openai/v1/route.test.ts
@@ -1134,6 +1134,7 @@ describe('v1CompletionsRoutes', () => {
       const app = createTestApp(
         createMockFluxService(),
         createMockConfigKV({
+          DEFAULT_TTS_MODEL: 'microsoft/v1',
           LLM_ROUTER_CONFIG: {
             llm: { models: {} },
             tts: {
@@ -1152,11 +1153,12 @@ describe('v1CompletionsRoutes', () => {
       )
 
       expect(res.status).toBe(200)
-      const data = await res.json() as { models: { id: string, name: string }[] }
+      const data = await res.json() as { models: { id: string, name: string }[], default: string }
       expect(data.models.map(m => m.id)).toEqual([
         'alibaba/cosyvoice-v2',
         'microsoft/v1',
       ])
+      expect(data.default).toBe('microsoft/v1')
     })
 
     it('returns an empty list when no tts models are configured', async () => {
diff --git a/packages/stage-ui/src/libs/providers/providers/official/index.ts b/packages/stage-ui/src/libs/providers/providers/official/index.ts
index 2494b685b..f4bc0d2c4 100644
--- a/packages/stage-ui/src/libs/providers/providers/official/index.ts
+++ b/packages/stage-ui/src/libs/providers/providers/official/index.ts
@@ -22,6 +22,15 @@ export const OFFICIAL_TRANSCRIPTION_PROVIDER_ID = 'official-provider-transcripti
 // first-voice matching when the server returns no recommendations.
 const recommendedVoicesByProvider: Record<string, Record<string, string>> = {}
 
+// Server-curated default HTTP speech model id, populated by the HTTP speech
+// provider's listModels(). The speech store uses this when it needs to seed an
+// empty/stale model selection, so the UI mirrors `/audio/speech` `model: auto`.
+let defaultSpeechModelId: string | null = null
+
+export function getDefaultSpeechModel(): string | null {
+  return defaultSpeechModelId
+}
+
 // Server-curated default streaming model id, populated by the streaming
 // provider's listModels(). Pages that need to seed an initial model selection
 // read this via getDefaultStreamingModel() instead of hardcoding an id.
@@ -117,14 +126,17 @@ export const providerOfficialSpeech = defineProvider({
   validationRequiredWhen: () => false,
   extraMethods: {
     listModels: async (): Promise<ModelInfo[]> => {
+      defaultSpeechModelId = null
       const res = await globalThis.fetch(`${SERVER_URL}/api/v1/audio/models`, { headers: authHeaders() })
       if (!res.ok)
         throw new Error(`audio models upstream ${res.status}: ${await res.text().catch(() => '')}`.slice(0, 256))
 
-      const data = await res.json() as { models?: { id: string, name: string }[] }
+      const data = await res.json() as { models?: { id: string, name: string }[], default?: string | null }
       if (!Array.isArray(data.models))
         throw new Error('audio models upstream returned malformed body')
 
+      defaultSpeechModelId = typeof data.default === 'string' && data.default.length > 0 ? data.default : null
+
       return data.models.map(m => ({
         id: m.id,
         name: m.name,
diff --git a/packages/stage-ui/src/stores/modules/speech.test.ts b/packages/stage-ui/src/stores/modules/speech.test.ts
index efdcc0a7a..07ac9e311 100644
--- a/packages/stage-ui/src/stores/modules/speech.test.ts
+++ b/packages/stage-ui/src/stores/modules/speech.test.ts
@@ -1,7 +1,7 @@
 import { createPinia, setActivePinia } from 'pinia'
 import { beforeEach, describe, expect, it, vi } from 'vitest'
 
-import { OFFICIAL_SPEECH_PROVIDER_ID, OFFICIAL_SPEECH_STREAMING_PROVIDER_ID } from '../../libs/providers/providers/official'
+import { OFFICIAL_SPEECH_PROVIDER_ID, OFFICIAL_SPEECH_STREAMING_PROVIDER_ID, providerOfficialSpeech } from '../../libs/providers/providers/official'
 import { useProvidersStore } from '../providers'
 import { toSignedPercent, useSpeechStore, voicePackForSpeechProvider } from './speech'
 
@@ -289,7 +289,29 @@ describe('speech store helpers', () => {
    * @example
    * speechStore.ensureActiveSpeechModel()
    */
-  it('resets stale streaming model when the regular official speech provider is active', () => {
+  it('resets stale streaming model to the server default when the regular official speech provider is active', async () => {
+    vi.stubGlobal('localStorage', {
+      getItem: vi.fn(() => null),
+      setItem: vi.fn(),
+      removeItem: vi.fn(),
+    })
+    vi.stubGlobal('fetch', vi.fn(async (input: RequestInfo | URL) => {
+      const url = input.toString()
+      if (url.includes('/api/v1/audio/models')) {
+        return new Response(JSON.stringify({
+          models: [
+            { id: 'alibaba/cosyvoice-v2', name: 'alibaba/cosyvoice-v2' },
+            { id: 'microsoft/v1', name: 'microsoft/v1' },
+          ],
+          default: 'microsoft/v1',
+        }), { status: 200, headers: { 'Content-Type': 'application/json' } })
+      }
+      return new Response(JSON.stringify({ voices: [], recommended: {} }), {
+        status: 200,
+        headers: { 'Content-Type': 'application/json' },
+      })
+    }) as typeof fetch)
+
     const providersStore = useProvidersStore()
     const speechStore = useSpeechStore()
     speechStore.activeSpeechProvider = OFFICIAL_SPEECH_PROVIDER_ID
@@ -301,15 +323,20 @@ describe('speech store helpers', () => {
       provider: OFFICIAL_SPEECH_STREAMING_PROVIDER_ID,
       languages: [],
     }
-    providersStore.providerRuntimeState[OFFICIAL_SPEECH_PROVIDER_ID].models = [
-      { id: 'microsoft/v1', name: 'microsoft/v1', provider: OFFICIAL_SPEECH_PROVIDER_ID },
-      { id: 'alibaba/cosyvoice-v2', name: 'alibaba/cosyvoice-v2', provider: OFFICIAL_SPEECH_PROVIDER_ID },
-    ]
+    try {
+      providersStore.providerRuntimeState[OFFICIAL_SPEECH_PROVIDER_ID].models = await providerOfficialSpeech.extraMethods!.listModels!(
+        {},
+        providerOfficialSpeech.createProvider({}),
+      )
 
-    speechStore.ensureActiveSpeechModel()
+      speechStore.ensureActiveSpeechModel()
 
-    expect(speechStore.activeSpeechModel).toBe('microsoft/v1')
-    expect(speechStore.activeSpeechVoiceId).toBe('')
-    expect(speechStore.activeSpeechVoice).toBeUndefined()
+      expect(speechStore.activeSpeechModel).toBe('microsoft/v1')
+      expect(speechStore.activeSpeechVoiceId).toBe('')
+      expect(speechStore.activeSpeechVoice).toBeUndefined()
+    }
+    finally {
+      vi.unstubAllGlobals()
+    }
   })
 })
diff --git a/packages/stage-ui/src/stores/modules/speech.ts b/packages/stage-ui/src/stores/modules/speech.ts
index e71b1b13c..9af4e0c5e 100644
--- a/packages/stage-ui/src/stores/modules/speech.ts
+++ b/packages/stage-ui/src/stores/modules/speech.ts
@@ -13,7 +13,7 @@ import { useI18n } from 'vue-i18n'
 import { toXml } from 'xast-util-to-xml'
 import { x } from 'xastscript'
 
-import { getDefaultStreamingModel, OFFICIAL_SPEECH_PROVIDER_ID, OFFICIAL_SPEECH_STREAMING_PROVIDER_ID, setupOfficialSpeechAutoPick } from '../../libs/providers/providers/official'
+import { getDefaultSpeechModel, getDefaultStreamingModel, OFFICIAL_SPEECH_PROVIDER_ID, OFFICIAL_SPEECH_STREAMING_PROVIDER_ID, setupOfficialSpeechAutoPick } from '../../libs/providers/providers/official'
 import { useProvidersStore } from '../providers'
 
 export function toSignedPercent(value: number): string {
@@ -286,7 +286,10 @@ export const useSpeechStore = defineStore('speech', () => {
     if (hasValidSelection)
       return
 
-    activeSpeechModel.value = models[0]?.id ?? ''
+    const defaultModel = getDefaultSpeechModel()
+    activeSpeechModel.value = defaultModel && models.some(m => m.id === defaultModel)
+      ? defaultModel
+      : models[0]?.id ?? ''
     clearVoiceSelection()
   }