fix: update default voice from "angus" to "luna" in text-to-speech functionality

- Changed the default voice for caching and initialization in the text-to-speech module to "luna" to align with the new model specifications.
- Updated related test cases and mock responses to reflect the new default voice.
- Adjusted comments in the code to clarify the default voice behavior.

Author: kentcdodds

app/components/calls/__tests__/submit-recording-form.test.tsx

@@ -20,6 +20,11 @@ vi.mock('#app/utils/use-root-data.ts', () => ({
 	useRootData: () => mockUseRootData(),
 }))
 
+// In Vitest, the Vite macro plugin isn't installed, so mock the macro helper.
+vi.mock('vite-env-only/macros', () => ({
+	serverOnly$: (fn: unknown) => fn,
+}))
+
 import { RecordingForm } from '#app/routes/resources/calls/save.tsx'
 
 describe('RecordingForm', () => {

app/routes/resources/calls/__tests__/text-to-speech-cache.test.ts

@@ -85,7 +85,7 @@ describe('/resources/calls/text-to-speech cache', () => {
 
 		const req1 = makeRequest({
 			text: 'Hello from the cache test message.',
-			voice: 'asteria',
+			voice: 'luna',
 		})
 		const res1 = (await action({ request: req1 } as any)) as Response
 		expect(res1.ok).toBe(true)
@@ -98,7 +98,7 @@ describe('/resources/calls/text-to-speech cache', () => {
 		// Same content, different whitespace: should hit cache.
 		const req2 = makeRequest({
 			text: '  Hello   from  the  cache   test message.  ',
-			voice: 'asteria',
+			voice: 'luna',
 		})
 		const res2 = (await action({ request: req2 } as any)) as Response
 		expect(res2.ok).toBe(true)

app/routes/resources/calls/text-to-speech.tsx

@@ -116,8 +116,8 @@ export async function action({ request }: Route.ActionArgs) {
 	const normalizedQuestionText = normalizeTextForCache(questionText)
 	const speechText = withAiDisclosurePrefix(normalizedQuestionText)
 	const model = getEnv().CLOUDFLARE_AI_TEXT_TO_SPEECH_MODEL
-	// Aura defaults to "angus" when omitted; treat empty voice as that for caching.
-	const voiceForCache = voiceRaw || 'angus'
+	// aura-2-en defaults to "luna" when omitted; treat empty voice as that for caching.
+	const voiceForCache = voiceRaw || 'luna'
 	const cacheKeyPayload = JSON.stringify({
 		v: 2,
 		model,
@@ -310,7 +310,7 @@ export function CallKentTextToSpeech({
 	const questionCountdownId = `${questionId}-countdown`
 
 	const defaultVoice = (callKentTextToSpeechVoices[0]?.id ??
-		'asteria') as CallKentTextToSpeechVoice
+		'luna') as CallKentTextToSpeechVoice
 	const [voice, setVoice] =
 		React.useState<CallKentTextToSpeechVoice>(defaultVoice)
 	const [questionText, setQuestionText] = React.useState('')

app/utils/call-kent-text-to-speech.ts

@@ -10,32 +10,25 @@ export const callKentTextToSpeechConstraints = {
 
 export const AI_VOICE_DISCLOSURE_PREFIX = `This caller's voice was generated by AI.`
 
+// @cf/deepgram/aura-2-en: 3 male + 3 female voices (Deepgram featured + model default)
 const auraSpeakers = [
-	'angus',
-	'asteria',
+	'apollo',
 	'arcas',
 	'orion',
-	'orpheus',
-	'athena',
 	'luna',
-	'zeus',
-	'perseus',
-	'helios',
-	'hera',
-	'stella',
+	'andromeda',
+	'helena',
 ] as const
 
 export type CallKentTextToSpeechVoice = (typeof auraSpeakers)[number]
 
 export const callKentTextToSpeechVoices: Array<{
 	id: CallKentTextToSpeechVoice
 	label: string
-}> = [
-	{ id: 'asteria', label: 'Asteria' },
-	{ id: 'orion', label: 'Orion' },
-	{ id: 'athena', label: 'Athena' },
-	{ id: 'angus', label: 'Angus' },
-]
+}> = auraSpeakers.map((speaker) => ({
+	id: speaker,
+	label: speaker.charAt(0).toUpperCase() + speaker.slice(1),
+}))
 
 export function isCallKentTextToSpeechVoice(
 	voice: string,

app/utils/cloudflare-ai-text-to-speech.server.ts

@@ -65,7 +65,7 @@ export async function synthesizeSpeechWithWorkersAi({
 		lowerModel.includes('deepgram/aura') || lowerModel.includes('aura-')
 			? {
 					text,
-					// Aura defaults to "angus", but allow callers to override.
+					// aura-2-en defaults to "luna", but allow callers to override.
 					...(voice ? { speaker: voice } : {}),
 					encoding: 'mp3',
 				}

mocks/__tests__/cloudflare.test.ts

@@ -72,7 +72,7 @@ describe('cloudflare MSW mocks', () => {
 				},
 				body: JSON.stringify({
 					text: 'Hello from the Call Kent TTS mock.',
-					speaker: 'asteria',
+					speaker: 'luna',
 					encoding: 'mp3',
 				}),
 			},

mocks/cloudflare.ts

@@ -943,7 +943,7 @@ export const cloudflareHandlers: Array<HttpHandler> = [
 				const speaker =
 					typeof body?.speaker === 'string' && body.speaker.trim()
 						? body.speaker.trim()
-						: 'angus'
+						: 'luna'
 				const frequencyHz = frequencyFromSpeakerAndText(speaker, text)
 				const wav = makePcm16SineWaveWav({
 					durationSeconds: 6,