genai-client-kit

dependencies {
    implementation(platform("io.github.ronjunevaldoz:genai-client-kit-bom:0.1.0"))
    implementation("io.github.ronjunevaldoz:genai-client-kit-elevenlabs")
}

dependencies {
    implementation("io.github.ronjunevaldoz:genai-client-kit-elevenlabs:0.1.0")
}

val client = ElevenLabsTranscriptionClient(apiKey = "...")

val result = client.transcribe(
    TranscriptionRequest(
        audio = audioBytes,
        filename = "clip.mp3",
        mimeType = "audio/mpeg",
        model = ElevenLabsModels.SCRIBE_V1, // or SCRIBE_V2
    ),
)
println(result.text)

val client = ElevenLabsSpeechClient(apiKey = "...")

val audioBytes = client.synthesizeSpeech(
    SpeechRequest(input = "Hello there", model = ElevenLabsModels.ELEVEN_TURBO_V2_5, voice = "voice-id"),
)

// or stream chunks as they arrive:
client.synthesizeSpeechStream(request).collect { chunk -> play(chunk) }

val realtime = ElevenLabsRealtimeClient(apiKey = "...")
val session = realtime.connect(voiceId = "voice-id")

launch {
    session.events.collect { event ->
        if (event is RealtimeEvent.Audio) play(event.bytes)
    }
}

session.send(RealtimeInputChunk.Text("Streaming this "))
session.send(RealtimeInputChunk.Text("as it's typed."))
session.flush()
session.close()

val voices = ElevenLabsVoicesClient(apiKey = "...")

val library = voices.listVoices()
val cloned = voices.addVoice(name = "My Voice", samples = listOf(VoiceSample("sample.mp3", "audio/mpeg", audioBytes)))
voices.editVoiceSettings(cloned.voiceId, VoiceSettings(stability = 0.6, similarityBoost = 0.8))

val agents = ElevenLabsAgentsClient(apiKey = "...")
val agent = agents.createAgent(
    name = "Support Bot",
    conversationConfig = buildJsonObject { put("first_message", "Hi, how can I help?") },
)

val conversations = ElevenLabsConversationsClient(apiKey = "...")
val page = conversations.listConversations(agentId = agent.agentId)
val detail = conversations.getConversation(page.conversations.first().conversationId)

val sts = ElevenLabsSpeechToSpeechClient(apiKey = "...")
val revoiced = sts.convert(SpeechToSpeechRequest(audio = clipBytes, filename = "clip.mp3", mimeType = "audio/mpeg", voiceId = "voice-id", model = ElevenLabsModels.ELEVEN_MULTILINGUAL_V2))

val effects = ElevenLabsSoundEffectsClient(apiKey = "...")
val doorCreak = effects.generate(SoundEffectRequest(text = "a door creaking open"))

val isolation = ElevenLabsAudioIsolationClient(apiKey = "...")
val voiceOnly = isolation.isolate(AudioIsolationRequest(audio = noisyClipBytes, filename = "clip.mp3", mimeType = "audio/mpeg"))

val dubbing = ElevenLabsDubbingClient(apiKey = "...")
val job = dubbing.createDubbing(DubbingSource("clip.mp4", "video/mp4", videoBytes), targetLanguage = "es")
// poll dubbing.getDubbingStatus(job.dubbingId) until status == "dubbed", then:
val dubbedAudio = dubbing.getDubbedAudio(job.dubbingId, languageCode = "es")

val knowledgeBase = ElevenLabsKnowledgeBaseClient(apiKey = "...")
val doc = knowledgeBase.createFromText(name = "FAQ", text = "Q: ... A: ...")

val agents = ElevenLabsAgentsClient(apiKey = "...")
val currentConfig = agents.getAgent(agent.agentId)["conversation_config"]!!.jsonObject
agents.updateAgent(agent.agentId, currentConfig.withKnowledgeBaseDocument(doc))

val calls = ElevenLabsOutboundCallClient(apiKey = "...")
val phoneNumber = calls.listPhoneNumbers().first()
calls.placeCall(agentId = agent.agentId, agentPhoneNumberId = phoneNumber.phoneNumberId, toNumber = "+15551234567")

val models = ElevenLabsModelsClient(apiKey = "...").listModels()

val user = ElevenLabsUserClient(apiKey = "...")
val subscription = user.getSubscription()
println("${subscription.characterCount}/${subscription.characterLimit} characters used")

val textToVoice = ElevenLabsTextToVoiceClient(apiKey = "...")
val previews = textToVoice.createPreviews(voiceDescription = "A calm, warm narrator")
val newVoice = textToVoice.createVoiceFromPreview("Narrator", "A calm, warm narrator", previews.first())

val library = ElevenLabsVoiceLibraryClient(apiKey = "...")
val results = library.search(language = "en", gender = "female")
library.addSharedVoice(results.voices.first().publicOwnerId, results.voices.first().voiceId, newName = "My Copy")

val history = ElevenLabsHistoryClient(apiKey = "...")
val page = history.listHistory()
val audio = history.getHistoryItemAudio(page.items.first().historyItemId)

val alignment = ElevenLabsForcedAlignmentClient(apiKey = "...")
val timings = alignment.align(audio = clipBytes, filename = "clip.mp3", mimeType = "audio/mpeg", text = "Hello there")

val music = ElevenLabsMusicClient(apiKey = "...")
val plan = music.createCompositionPlan(prompt = "An upbeat synthwave track", musicLengthMs = 30000)
val clip = music.compose(compositionPlan = plan)
// or straight from a prompt, skipping the plan step:
val quickClip = music.compose(prompt = "A calm lo-fi beat", musicLengthMs = 15000)

val dictionaries = ElevenLabsPronunciationDictionariesClient(apiKey = "...")
val dictionary = dictionaries.createFromRules(
    name = "Product Names",
    rules = listOf(PronunciationRule("Kubernetes", PronunciationRuleType.ALIAS, alias = "koo-ber-NET-eez")),
)
// then pass dictionary.id into a TTS/dubbing request via provider-specific options once supported

val projects = ElevenLabsProjectsClient(apiKey = "...")
val project = projects.createProject(name = "My Audiobook", defaultParagraphVoiceId = "voice-id")
projects.convertProject(project.projectId)
val chapters = projects.listChapters(project.projectId)
val snapshots = projects.listChapterSnapshots(project.projectId, chapters.first().chapterId)
val audio = projects.getChapterSnapshotAudio(project.projectId, chapters.first().chapterId, snapshots.first().snapshotId)

val webhooks = ElevenLabsWebhooksClient(apiKey = "...").listWebhooks()