{"id":"0d8c322d-0bbf-4077-8a98-79b4c09fcab0","shortId":"rzbSP5","kind":"skill","title":"voice-ai-development","tagline":"Expert in building voice AI applications - from real-time voice","description":"# Voice AI Development\n\nExpert in building voice AI applications - from real-time voice agents to voice-enabled apps.\nCovers OpenAI Realtime API, Vapi for voice agents, Deepgram for transcription, ElevenLabs\nfor synthesis, LiveKit for real-time infrastructure, and WebRTC fundamentals. Knows how to\nbuild low-latency, production-ready voice experiences.\n\n**Role**: Voice AI Architect\n\nYou are an expert in building real-time voice applications. You think in terms of\nlatency budgets, audio quality, and user experience. You know that voice apps feel\nmagical when fast and broken when slow. You choose the right combination of providers\nfor each use case and optimize relentlessly for perceived responsiveness.\n\n### Expertise\n\n- Real-time audio streaming\n- Voice agent architecture\n- Provider selection\n- Latency optimization\n- Audio quality tuning\n\n## Capabilities\n\n- OpenAI Realtime API\n- Vapi voice agents\n- Deepgram STT/TTS\n- ElevenLabs voice synthesis\n- LiveKit real-time infrastructure\n- WebRTC audio handling\n- Voice agent design\n- Latency optimization\n\n## Prerequisites\n\n- 0: Async programming\n- 1: WebSocket basics\n- 2: Audio concepts (sample rate, codec)\n- Required skills: Python or Node.js, API keys for providers, Audio handling knowledge\n\n## Scope\n\n- 0: Latency varies by provider\n- 1: Cost per minute adds up\n- 2: Quality depends on network\n- 3: Complex debugging\n\n## Ecosystem\n\n### Primary\n\n- OpenAI Realtime API\n- Vapi\n- Deepgram\n- ElevenLabs\n\n### Infrastructure\n\n- LiveKit\n- Daily.co\n- Twilio\n\n### Common_integrations\n\n- WebRTC\n- WebSockets\n- Telephony (SIP/PSTN)\n\n### Platforms\n\n- Web applications\n- Mobile apps\n- Call centers\n- Voice assistants\n\n## Patterns\n\n### OpenAI Realtime API\n\nNative voice-to-voice with GPT-4o\n\n**When to use**: When you want integrated voice AI without separate STT/TTS\n\nimport asyncio\nimport websockets\nimport json\nimport base64\n\nOPENAI_API_KEY = \"sk-...\"\n\nasync def voice_session():\n    url = \"wss://api.openai.com/v1/realtime?model=gpt-4o-realtime-preview\"\n    headers = {\n        \"Authorization\": f\"Bearer {OPENAI_API_KEY}\",\n        \"OpenAI-Beta\": \"realtime=v1\"\n    }\n\n    async with websockets.connect(url, extra_headers=headers) as ws:\n        # Configure session\n        await ws.send(json.dumps({\n            \"type\": \"session.update\",\n            \"session\": {\n                \"modalities\": [\"text\", \"audio\"],\n                \"voice\": \"alloy\",  # alloy, echo, fable, onyx, nova, shimmer\n                \"input_audio_format\": \"pcm16\",\n                \"output_audio_format\": \"pcm16\",\n                \"input_audio_transcription\": {\n                    \"model\": \"whisper-1\"\n                },\n                \"turn_detection\": {\n                    \"type\": \"server_vad\",  # Voice activity detection\n                    \"threshold\": 0.5,\n                    \"prefix_padding_ms\": 300,\n                    \"silence_duration_ms\": 500\n                },\n                \"tools\": [\n                    {\n                        \"type\": \"function\",\n                        \"name\": \"get_weather\",\n                        \"description\": \"Get weather for a location\",\n                        \"parameters\": {\n                            \"type\": \"object\",\n                            \"properties\": {\n                                \"location\": {\"type\": \"string\"}\n                            }\n                        }\n                    }\n                ]\n            }\n        }))\n\n        # Send audio (PCM16, 24kHz, mono)\n        async def send_audio(audio_bytes):\n            await ws.send(json.dumps({\n                \"type\": \"input_audio_buffer.append\",\n                \"audio\": base64.b64encode(audio_bytes).decode()\n            }))\n\n        # Receive events\n        async for message in ws:\n            event = json.loads(message)\n\n            if event[\"type\"] == \"response.audio.delta\":\n                # Play audio chunk\n                audio = base64.b64decode(event[\"delta\"])\n                play_audio(audio)\n\n            elif event[\"type\"] == \"response.audio_transcript.done\":\n                print(f\"Assistant said: {event['transcript']}\")\n\n            elif event[\"type\"] == \"input_audio_buffer.speech_started\":\n                print(\"User started speaking\")\n\n            elif event[\"type\"] == \"response.function_call_arguments.done\":\n                # Handle tool call\n                name = event[\"name\"]\n                args = json.loads(event[\"arguments\"])\n                result = call_function(name, args)\n                await ws.send(json.dumps({\n                    \"type\": \"conversation.item.create\",\n                    \"item\": {\n                        \"type\": \"function_call_output\",\n                        \"call_id\": event[\"call_id\"],\n                        \"output\": json.dumps(result)\n                    }\n                }))\n\n### Vapi Voice Agent\n\nBuild voice agents with Vapi platform\n\n**When to use**: Phone-based agents, quick deployment\n\n# Vapi provides hosted voice agents with webhooks\n\nfrom flask import Flask, request, jsonify\nimport vapi\n\napp = Flask(__name__)\nclient = vapi.Vapi(api_key=\"...\")\n\n# Create an assistant\nassistant = client.assistants.create(\n    name=\"Support Agent\",\n    model={\n        \"provider\": \"openai\",\n        \"model\": \"gpt-4o\",\n        \"messages\": [\n            {\n                \"role\": \"system\",\n                \"content\": \"You are a helpful support agent...\"\n            }\n        ]\n    },\n    voice={\n        \"provider\": \"11labs\",\n        \"voiceId\": \"21m00Tcm4TlvDq8ikWAM\"  # Rachel\n    },\n    firstMessage=\"Hi! How can I help you today?\",\n    transcriber={\n        \"provider\": \"deepgram\",\n        \"model\": \"nova-2\"\n    }\n)\n\n# Webhook for conversation events\n@app.route(\"/vapi/webhook\", methods=[\"POST\"])\ndef vapi_webhook():\n    event = request.json\n\n    if event[\"type\"] == \"function-call\":\n        # Handle tool call\n        name = event[\"functionCall\"][\"name\"]\n        args = event[\"functionCall\"][\"parameters\"]\n\n        if name == \"check_order\":\n            result = check_order(args[\"order_id\"])\n            return jsonify({\"result\": result})\n\n    elif event[\"type\"] == \"end-of-call-report\":\n        # Call ended - save transcript\n        transcript = event[\"transcript\"]\n        save_transcript(event[\"call\"][\"id\"], transcript)\n\n    return jsonify({\"ok\": True})\n\n# Start outbound call\ncall = client.calls.create(\n    assistant_id=assistant.id,\n    customer={\n        \"number\": \"+1234567890\"\n    },\n    phoneNumber={\n        \"twilioPhoneNumber\": \"+0987654321\"\n    }\n)\n\n# Or create web call\nweb_call = client.calls.create(\n    assistant_id=assistant.id,\n    type=\"web\"\n)\n# Returns URL for WebRTC connection\n\n### Deepgram STT + ElevenLabs TTS\n\nBest-in-class transcription and synthesis\n\n**When to use**: High quality voice, custom pipeline\n\nimport asyncio\nfrom deepgram import DeepgramClient, LiveTranscriptionEvents\nfrom elevenlabs import ElevenLabs\n\n# Deepgram real-time transcription\ndeepgram = DeepgramClient(api_key=\"...\")\n\nasync def transcribe_stream(audio_stream):\n    connection = deepgram.listen.live.v(\"1\")\n\n    async def on_transcript(result):\n        transcript = result.channel.alternatives[0].transcript\n        if transcript:\n            print(f\"Heard: {transcript}\")\n            if result.is_final:\n                # Process final transcript\n                await handle_user_input(transcript)\n\n    connection.on(LiveTranscriptionEvents.Transcript, on_transcript)\n\n    await connection.start({\n        \"model\": \"nova-2\",  # Best quality\n        \"language\": \"en\",\n        \"smart_format\": True,\n        \"interim_results\": True,  # Get partial results\n        \"utterance_end_ms\": 1000,\n        \"vad_events\": True,  # Voice activity detection\n        \"encoding\": \"linear16\",\n        \"sample_rate\": 16000\n    })\n\n    # Stream audio\n    async for chunk in audio_stream:\n        await connection.send(chunk)\n\n    await connection.finish()\n\n# ElevenLabs streaming synthesis\neleven = ElevenLabs(api_key=\"...\")\n\ndef text_to_speech_stream(text: str):\n    \"\"\"Stream TTS audio chunks.\"\"\"\n    audio_stream = eleven.text_to_speech.convert_as_stream(\n        voice_id=\"21m00Tcm4TlvDq8ikWAM\",  # Rachel\n        model_id=\"eleven_turbo_v2_5\",  # Fastest\n        text=text,\n        output_format=\"pcm_24000\"  # Raw PCM for low latency\n    )\n\n    for chunk in audio_stream:\n        yield chunk\n\n# Or with WebSocket for lowest latency\nasync def tts_websocket(text_stream):\n    async with eleven.text_to_speech.stream_async(\n        voice_id=\"21m00Tcm4TlvDq8ikWAM\",\n        model_id=\"eleven_turbo_v2_5\"\n    ) as tts:\n        async for text_chunk in text_stream:\n            audio = await tts.send(text_chunk)\n            yield audio\n\n        # Flush remaining audio\n        final_audio = await tts.flush()\n        yield final_audio\n\n### LiveKit Real-time Infrastructure\n\nWebRTC infrastructure for voice apps\n\n**When to use**: Building custom real-time voice apps\n\nfrom livekit import api, rtc\nimport asyncio\n\n# Server-side: Create room and tokens\nlk_api = api.LiveKitAPI(\n    url=\"wss://your-livekit.livekit.cloud\",\n    api_key=\"...\",\n    api_secret=\"...\"\n)\n\nasync def create_room(room_name: str):\n    room = await lk_api.room.create_room(\n        api.CreateRoomRequest(name=room_name)\n    )\n    return room\n\ndef create_token(room_name: str, participant_name: str):\n    token = api.AccessToken(\n        api_key=\"...\",\n        api_secret=\"...\"\n    )\n    token.with_identity(participant_name)\n    token.with_grants(api.VideoGrants(\n        room_join=True,\n        room=room_name\n    ))\n    return token.to_jwt()\n\n# Agent-side: Connect and process audio\nasync def voice_agent(room_name: str):\n    room = rtc.Room()\n\n    @room.on(\"track_subscribed\")\n    def on_track(track, publication, participant):\n        if track.kind == rtc.TrackKind.KIND_AUDIO:\n            # Process incoming audio\n            audio_stream = rtc.AudioStream(track)\n            asyncio.create_task(process_audio(audio_stream))\n\n    token = create_token(room_name, \"agent\")\n    await room.connect(\"wss://your-livekit.livekit.cloud\", token)\n\n    # Publish agent's audio\n    source = rtc.AudioSource(sample_rate=24000, num_channels=1)\n    track = rtc.LocalAudioTrack.create_audio_track(\"agent-voice\", source)\n    await room.local_participant.publish_track(track)\n\n    # Send audio from TTS\n    async def speak(text: str):\n        for audio_chunk in text_to_speech(text):\n            await source.capture_frame(rtc.AudioFrame(\n                data=audio_chunk,\n                sample_rate=24000,\n                num_channels=1,\n                samples_per_channel=len(audio_chunk) // 2\n            ))\n\n    return room, speak\n\n# Process audio with STT\nasync def process_audio(audio_stream):\n    async for frame in audio_stream:\n        # Send to Deepgram or other STT\n        await transcriber.send(frame.data)\n\n### Full Voice Agent Pipeline\n\nComplete voice agent with all components\n\n**When to use**: Custom production voice agent\n\nimport asyncio\nfrom dataclasses import dataclass\nfrom typing import AsyncIterator\n\n@dataclass\nclass VoiceAgentConfig:\n    stt_provider: str = \"deepgram\"\n    tts_provider: str = \"elevenlabs\"\n    llm_provider: str = \"openai\"\n    vad_enabled: bool = True\n    interrupt_enabled: bool = True\n\nclass VoiceAgent:\n    def __init__(self, config: VoiceAgentConfig):\n        self.config = config\n        self.is_speaking = False\n        self.conversation_history = []\n\n    async def process_audio_stream(\n        self,\n        audio_in: AsyncIterator[bytes],\n        audio_out: asyncio.Queue\n    ):\n        \"\"\"Main audio processing loop.\"\"\"\n\n        # STT streaming\n        async def transcribe():\n            transcript_buffer = \"\"\n            async for audio_chunk in audio_in:\n                # Check for interruption\n                if self.is_speaking and self.config.interrupt_enabled:\n                    if await self.detect_speech(audio_chunk):\n                        await self.stop_speaking()\n\n                result = await self.stt.transcribe(audio_chunk)\n                if result.is_final:\n                    yield result.transcript\n\n        # Process transcripts\n        async for user_text in transcribe():\n            if not user_text.strip():\n                continue\n\n            self.conversation_history.append({\n                \"role\": \"user\",\n                \"content\": user_text\n            })\n\n            # Generate response with streaming\n            self.is_speaking = True\n            async for audio_chunk in self.generate_response(user_text):\n                await audio_out.put(audio_chunk)\n            self.is_speaking = False\n\n    async def generate_response(self, text: str) -> AsyncIterator[bytes]:\n        \"\"\"Stream LLM response through TTS.\"\"\"\n\n        # Stream LLM tokens\n        llm_stream = self.llm.stream_chat(self.conversation_history)\n\n        # Buffer for TTS (need ~50 chars for good prosody)\n        text_buffer = \"\"\n        full_response = \"\"\n\n        async for token in llm_stream:\n            text_buffer += token\n            full_response += token\n\n            # Send to TTS when we have enough text\n            if len(text_buffer) > 50 or token in \".!?\":\n                async for audio in self.tts.synthesize_stream(text_buffer):\n                    yield audio\n                text_buffer = \"\"\n\n        # Flush remaining\n        if text_buffer:\n            async for audio in self.tts.synthesize_stream(text_buffer):\n                yield audio\n\n        self.conversation_history.append({\n            \"role\": \"assistant\",\n            \"content\": full_response\n        })\n\n    async def detect_speech(self, audio: bytes) -> bool:\n        \"\"\"Voice activity detection.\"\"\"\n        # Use WebRTC VAD or Silero VAD\n        return self.vad.is_speech(audio)\n\n    async def stop_speaking(self):\n        \"\"\"Handle interruption.\"\"\"\n        self.is_speaking = False\n        # Clear audio queue\n        # Stop TTS generation\n\n# Latency optimization tips:\n# 1. Use streaming everywhere (STT, LLM, TTS)\n# 2. Start TTS before LLM finishes (~50 char buffer)\n# 3. Use PCM audio format (no encoding overhead)\n# 4. Keep WebSocket connections alive\n# 5. Use regional endpoints close to users\n\n## Validation Checks\n\n### Non-Streaming TTS\n\nSeverity: HIGH\n\nMessage: Non-streaming TTS adds significant latency.\n\nFix action: Use tts.synthesize_stream() or tts.convert_as_stream()\n\n### Hardcoded Sample Rate\n\nSeverity: MEDIUM\n\nMessage: Hardcoded sample rate may cause format mismatches.\n\nFix action: Define sample rates as constants, document expected formats\n\n### WebSocket Without Reconnection\n\nSeverity: HIGH\n\nMessage: WebSocket connections need reconnection logic.\n\nFix action: Add retry loop with exponential backoff\n\n### Missing VAD Configuration\n\nSeverity: MEDIUM\n\nMessage: VAD needs tuning for good user experience.\n\nFix action: Configure threshold and silence_duration_ms\n\n### Blocking Audio Processing\n\nSeverity: HIGH\n\nMessage: Audio processing should be async to avoid blocking.\n\nFix action: Use async def and await for audio operations\n\n### Missing Interruption Handling\n\nSeverity: MEDIUM\n\nMessage: Voice agents should handle user interruptions.\n\nFix action: Add barge-in detection and cancel current response\n\n### Audio Queue Without Clear\n\nSeverity: LOW\n\nMessage: Audio queues should be clearable for interruptions.\n\nFix action: Add method to clear queue on interruption\n\n### WebSocket Without Error Handling\n\nSeverity: HIGH\n\nMessage: WebSocket operations need error handling.\n\nFix action: Wrap in try/except for ConnectionClosed\n\n## Collaboration\n\n### Delegation Triggers\n\n- agent graph|workflow|state -> langgraph (Need complex agent logic behind voice)\n- extract|structured|json -> structured-output (Need to extract structured data from voice)\n- observability|tracing|monitoring -> langfuse (Need to monitor voice agent quality)\n- frontend|web|react -> nextjs-app-router (Need web interface for voice agent)\n\n### Intelligent Voice Agent\n\nSkills: voice-ai-development, langgraph, structured-output\n\nWorkflow:\n\n```\n1. Design agent graph with tools\n2. Add voice interface layer\n3. Use structured output for tool responses\n4. Optimize for voice latency\n```\n\n### Monitored Voice Agent\n\nSkills: voice-ai-development, langfuse\n\nWorkflow:\n\n```\n1. Build voice agent with provider of choice\n2. Add Langfuse callbacks\n3. Track latency, quality, conversation flow\n4. Iterate based on metrics\n```\n\n### Phone-based Agent\n\nSkills: voice-ai-development, twilio\n\nWorkflow:\n\n```\n1. Set up Vapi or custom agent\n2. Connect to Twilio for PSTN\n3. Handle inbound/outbound calls\n4. Implement call routing logic\n```\n\n## Related Skills\n\nWorks well with: `langgraph`, `structured-output`, `langfuse`\n\n## When to Use\n- User mentions or implies: voice ai\n- User mentions or implies: voice agent\n- User mentions or implies: speech to text\n- User mentions or implies: text to speech\n- User mentions or implies: realtime voice\n- User mentions or implies: vapi\n- User mentions or implies: deepgram\n- User mentions or implies: elevenlabs\n- User mentions or implies: livekit\n- User mentions or implies: openai realtime\n\n## Limitations\n- Use this skill only when the task clearly matches the scope described above.\n- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.\n- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.","tags":["voice","development","antigravity","awesome","skills","sickn33","agent-skills","agentic-skills","ai-agent-skills","ai-agents","ai-coding","ai-workflows"],"capabilities":["skill","source-sickn33","skill-voice-ai-development","topic-agent-skills","topic-agentic-skills","topic-ai-agent-skills","topic-ai-agents","topic-ai-coding","topic-ai-workflows","topic-antigravity","topic-antigravity-skills","topic-claude-code","topic-claude-code-skills","topic-codex-cli","topic-codex-skills"],"categories":["antigravity-awesome-skills"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/sickn33/antigravity-awesome-skills/voice-ai-development","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add sickn33/antigravity-awesome-skills","source_repo":"https://github.com/sickn33/antigravity-awesome-skills","install_from":"skills.sh"}},"qualityScore":"0.700","qualityRationale":"deterministic score 0.70 from registry signals: · indexed on github topic:agent-skills · 34997 github stars · SKILL.md body (17,340 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-04-25T06:52:16.742Z","embedding":null,"createdAt":"2026-04-18T20:38:59.220Z","updatedAt":"2026-04-25T06:52:16.742Z","lastSeenAt":"2026-04-25T06:52:16.742Z","tsv":"'+0987654321':645 '+1234567890':642 '-1':339 '-2':562,746 '/v1/realtime?model=gpt-4o-realtime-preview':285 '/vapi/webhook':568 '0':170,195,719 '0.5':349 '1':173,200,711,1045,1087,1424,1700,1733,1767 '1000':763 '11labs':545 '16000':774 '2':176,206,1094,1431,1706,1741,1774 '21m00tcm4tlvdq8ikwam':547,813,858 '24000':827,1042,1084 '24khz':380 '3':211,1440,1711,1745,1780 '300':353 '4':1448,1718,1751,1784 '4o':253,532 '5':820,864,1453 '50':1314,1347,1437 '500':357 'action':1477,1499,1520,1541,1563,1585,1610,1631 'activ':346,768,1393 'add':204,1473,1521,1586,1611,1707,1742 'agent':30,43,135,150,165,480,483,493,500,525,542,983,992,1029,1035,1051,1125,1129,1139,1579,1640,1647,1672,1686,1689,1702,1725,1736,1759,1773,1813 'agent-sid':982 'agent-voic':1050 'ai':3,9,17,23,73,262,1693,1729,1763,1807 'aliv':1452 'alloy':319,320 'api':39,147,187,218,244,275,291,516,700,793,914,926,930,932,962,964 'api.accesstoken':961 'api.createroomrequest':945 'api.livekitapi':927 'api.openai.com':284 'api.openai.com/v1/realtime?model=gpt-4o-realtime-preview':283 'api.videogrants':972 'app':35,102,236,511,900,910,1679 'app.route':567 'applic':10,24,85,234 'architect':74 'architectur':136 'arg':451,459,589,600 'argument':454 'ask':1893 'assist':240,428,520,521,637,653,1380 'assistant.id':639,655 'async':171,278,298,382,400,702,712,777,846,852,855,867,934,989,1062,1102,1108,1187,1206,1211,1248,1271,1287,1323,1351,1368,1384,1405,1558,1565 'asyncio':267,683,917,1141 'asyncio.create':1018 'asyncio.queue':1199 'asynciter':1149,1195,1294 'audio':93,132,141,162,177,191,317,327,331,335,378,385,386,393,395,413,415,420,421,706,776,781,804,806,836,874,880,883,885,890,988,1010,1013,1014,1021,1022,1037,1048,1059,1068,1080,1092,1099,1105,1106,1112,1190,1193,1197,1201,1213,1216,1231,1239,1273,1282,1353,1360,1370,1377,1389,1404,1416,1443,1549,1554,1570,1595,1602 'audio_out.put':1281 'author':287 'avoid':1560 'await':309,388,460,733,742,783,786,875,886,942,1030,1054,1075,1120,1228,1233,1237,1280,1568 'backoff':1526 'barg':1588 'barge-in':1587 'base':492,1753,1758 'base64':273 'base64.b64decode':416 'base64.b64encode':394 'basic':175 'bearer':289 'behind':1649 'best':668,747 'best-in-class':667 'beta':295 'block':1548,1561 'bool':1167,1171,1391 'boundari':1901 'broken':108 'budget':92 'buffer':1210,1310,1320,1330,1346,1358,1362,1367,1375,1439 'build':7,21,62,80,481,904,1734 'byte':387,396,1196,1295,1390 'call':237,447,456,468,470,473,581,584,613,615,625,634,635,649,651,1783,1786 'callback':1744 'cancel':1592 'capabl':144 'case':121 'caus':1495 'center':238 'channel':1044,1086,1090 'char':1315,1438 'chat':1307 'check':595,598,1218,1461 'choic':1740 'choos':112 'chunk':414,779,785,805,834,839,870,878,1069,1081,1093,1214,1232,1240,1274,1283 'clarif':1895 'class':670,1151,1173 'clear':1415,1598,1614,1868 'clearabl':1606 'client':514 'client.assistants.create':522 'client.calls.create':636,652 'close':1457 'codec':181 'collabor':1637 'combin':115 'common':226 'complet':1127 'complex':212,1646 'compon':1132 'concept':178 'config':1178,1181 'configur':307,1529,1542 'connect':662,708,985,1451,1515,1775 'connection.finish':787 'connection.on':738 'connection.send':784 'connection.start':743 'connectionclos':1636 'constant':1504 'content':536,1261,1381 'continu':1257 'convers':565,1749 'conversation.item.create':464 'cost':201 'cover':36 'creat':518,647,921,936,952,1025 'criteria':1904 'current':1593 'custom':640,680,905,1136,1772 'daily.co':224 'data':1079,1661 'dataclass':1143,1145,1150 'debug':213 'decod':397 'deepgram':44,151,220,559,663,685,693,698,1116,1156,1843 'deepgram.listen.live':709 'deepgramcli':687,699 'def':279,383,571,703,713,795,847,935,951,990,1001,1063,1103,1175,1188,1207,1288,1385,1406,1566 'defin':1500 'deleg':1638 'delta':418 'depend':208 'deploy':495 'describ':1872 'descript':364 'design':166,1701 'detect':341,347,769,1386,1394,1590 'develop':4,18,1694,1730,1764 'document':1505 'durat':355,1546 'echo':321 'ecosystem':214 'eleven':791,817,861 'eleven.text_to_speech.convert':808 'eleven.text_to_speech.stream':854 'elevenlab':47,153,221,665,690,692,788,792,1160,1848 'elif':422,432,441,607 'en':750 'enabl':34,1166,1170,1226 'encod':770,1446 'end':611,616,761 'end-of-call-report':610 'endpoint':1456 'enough':1341 'environ':1884 'environment-specif':1883 'error':1620,1628 'event':399,405,409,417,423,430,433,442,449,453,472,566,574,577,586,590,608,620,624,765 'everywher':1427 'expect':1506 'experi':70,97,1539 'expert':5,19,78,1889 'expertis':128 'exponenti':1525 'extra':302 'extract':1651,1659 'f':288,427,724 'fabl':322 'fals':1184,1286,1414 'fast':106 'fastest':821 'feel':103 'final':729,731,884,889,1243 'finish':1436 'firstmessag':549 'fix':1476,1498,1519,1540,1562,1584,1609,1630 'flask':504,506,512 'flow':1750 'flush':881,1363 'format':328,332,752,825,1444,1496,1507 'frame':1077,1110 'frame.data':1122 'frontend':1674 'full':1123,1321,1332,1382 'function':360,457,467,580 'function-cal':579 'functioncal':587,591 'fundament':58 'generat':1264,1289,1420 'get':362,365,757 'good':1317,1537 'gpt':252,531 'gpt-4o':251,530 'grant':971 'graph':1641,1703 'handl':163,192,445,582,734,1410,1574,1581,1621,1629,1781 'hardcod':1485,1491 'header':286,303,304 'heard':725 'help':540,554 'hi':550 'high':677,1467,1512,1552,1623 'histori':1186,1309 'host':498 'id':471,474,602,626,638,654,812,816,857,860 'ident':967 'implement':1785 'impli':1805,1811,1817,1824,1831,1837,1842,1847,1852,1857 'import':266,268,270,272,505,509,682,686,691,913,916,1140,1144,1148 'inbound/outbound':1782 'incom':1012 'infrastructur':55,160,222,895,897 'init':1176 'input':326,334,736,1898 'input_audio_buffer.append':392 'input_audio_buffer.speech':435 'integr':227,260 'intellig':1687 'interfac':1683,1709 'interim':754 'interrupt':1169,1220,1411,1573,1583,1608,1617 'item':465 'iter':1752 'join':974 'json':271,1653 'json.dumps':311,390,462,476 'json.loads':406,452 'jsonifi':508,604,629 'jwt':981 'keep':1449 'key':188,276,292,517,701,794,931,963 'know':59,99 'knowledg':193 'langfus':1667,1731,1743,1798 'langgraph':1644,1695,1794 'languag':749 'latenc':65,91,139,167,196,832,845,1421,1475,1722,1747 'layer':1710 'len':1091,1344 'limit':1860 'linear16':771 'livekit':50,156,223,891,912,1853 'livetranscriptionev':688 'livetranscriptionevents.transcript':739 'lk':925 'lk_api.room.create':943 'llm':1161,1297,1302,1304,1327,1429,1435 'locat':369,374 'logic':1518,1648,1788 'loop':1203,1523 'low':64,831,1600 'low-lat':63 'lowest':844 'magic':104 'main':1200 'match':1869 'may':1494 'medium':1489,1531,1576 'mention':1803,1809,1815,1822,1829,1835,1840,1845,1850,1855 'messag':402,407,533,1468,1490,1513,1532,1553,1577,1601,1624 'method':569,1612 'metric':1755 'minut':203 'mismatch':1497 'miss':1527,1572,1906 'mobil':235 'modal':315 'model':337,526,529,560,744,815,859 'monitor':1666,1670,1723 'mono':381 'ms':352,356,762,1547 'name':361,448,450,458,513,523,585,588,594,939,946,948,955,958,969,978,994,1028 'nativ':245 'need':1313,1516,1534,1627,1645,1657,1668,1681 'network':210 'nextj':1678 'nextjs-app-rout':1677 'node.js':186 'non':1463,1470 'non-stream':1462,1469 'nova':324,561,745 'num':1043,1085 'number':641 'object':372 'observ':1664 'ok':630 'onyx':323 'openai':37,145,216,242,274,290,294,528,1164,1858 'openai-beta':293 'oper':1571,1626 'optim':123,140,168,1422,1719 'order':596,599,601 'outbound':633 'output':330,469,475,824,1656,1698,1714,1797,1878 'overhead':1447 'pad':351 'paramet':370,592 'partial':758 'particip':957,968,1006 'pattern':241 'pcm':826,829,1442 'pcm16':329,333,379 'per':202,1089 'perceiv':126 'permiss':1899 'phone':491,1757 'phone-bas':490,1756 'phonenumb':643 'pipelin':681,1126 'platform':232,486 'play':412,419 'post':570 'prefix':350 'prerequisit':169 'primari':215 'print':426,437,723 'process':730,987,1011,1020,1098,1104,1189,1202,1246,1550,1555 'product':67,1137 'production-readi':66 'program':172 'properti':373 'prosodi':1318 'provid':117,137,190,199,497,527,544,558,1154,1158,1162,1738 'pstn':1779 'public':1005 'publish':1034 'python':184 'qualiti':94,142,207,678,748,1673,1748 'queue':1417,1596,1603,1615 'quick':494 'rachel':548,814 'rate':180,773,1041,1083,1487,1493,1502 'raw':828 'react':1676 'readi':68 'real':13,27,53,82,130,158,695,893,907 'real-tim':12,26,52,81,129,157,694,892,906 'realtim':38,146,217,243,296,1832,1859 'receiv':398 'reconnect':1510,1517 'region':1455 'relat':1789 'relentless':124 'remain':882,1364 'report':614 'request':507 'request.json':575 'requir':182,1897 'respons':127,1265,1277,1290,1298,1322,1333,1383,1594,1717 'response.audio.delta':411 'response.audio_transcript.done':425 'response.function_call_arguments.done':444 'result':455,477,597,605,606,716,755,759,1236 'result.channel.alternatives':718 'result.is':728,1242 'result.transcript':1245 'retri':1522 'return':603,628,658,949,979,1095,1401 'review':1890 'right':114 'role':71,534,1259,1379 'room':922,937,938,941,944,947,950,954,973,976,977,993,996,1027,1096 'room.connect':1031 'room.local_participant.publish':1055 'room.on':998 'rout':1787 'router':1680 'rtc':915 'rtc.audioframe':1078 'rtc.audiosource':1039 'rtc.audiostream':1016 'rtc.localaudiotrack.create':1047 'rtc.room':997 'rtc.trackkind.kind':1009 'safeti':1900 'said':429 'sampl':179,772,1040,1082,1088,1486,1492,1501 'save':617,622 'scope':194,1871 'secret':933,965 'select':138 'self':1177,1192,1291,1388,1409 'self.config':1180 'self.config.interrupt':1225 'self.conversation':1185,1308 'self.conversation_history.append':1258,1378 'self.detect':1229 'self.generate':1276 'self.is':1182,1222,1268,1284,1412 'self.llm.stream':1306 'self.stop':1234 'self.stt.transcribe':1238 'self.tts.synthesize':1355,1372 'self.vad.is':1402 'send':377,384,1058,1114,1335 'separ':264 'server':343,919 'server-sid':918 'session':281,308,314 'session.update':313 'set':1768 'sever':1466,1488,1511,1530,1551,1575,1599,1622 'shimmer':325 'side':920,984 'signific':1474 'silenc':354,1545 'silero':1399 'sip/pstn':231 'sk':277 'skill':183,1690,1726,1760,1790,1863 'skill-voice-ai-development' 'slow':110 'smart':751 'sourc':1038,1053 'source-sickn33' 'source.capture':1076 'speak':440,1064,1097,1183,1223,1235,1269,1285,1408,1413 'specif':1885 'speech':798,1073,1230,1387,1403,1818,1827 'start':436,439,632,1432 'state':1643 'stop':1407,1418,1891 'str':801,940,956,959,995,1066,1155,1159,1163,1293 'stream':133,705,707,775,782,789,799,802,807,810,837,851,873,1015,1023,1107,1113,1191,1205,1267,1296,1301,1305,1328,1356,1373,1426,1464,1471,1480,1484 'string':376 'structur':1652,1655,1660,1697,1713,1796 'structured-output':1654,1696,1795 'stt':664,1101,1119,1153,1204,1428 'stt/tts':152,265 'subscrib':1000 'substitut':1881 'success':1903 'support':524,541 'synthesi':49,155,673,790 'system':535 'task':1019,1867 'telephoni':230 'term':89 'test':1887 'text':316,796,800,822,823,850,869,872,877,1065,1071,1074,1251,1263,1279,1292,1319,1329,1342,1345,1357,1361,1366,1374,1820,1825 'think':87 'threshold':348,1543 'time':14,28,54,83,131,159,696,894,908 'tip':1423 'today':556 'token':924,953,960,1024,1026,1033,1303,1325,1331,1334,1349 'token.to':980 'token.with':966,970 'tool':358,446,583,1705,1716 'topic-agent-skills' 'topic-agentic-skills' 'topic-ai-agent-skills' 'topic-ai-agents' 'topic-ai-coding' 'topic-ai-workflows' 'topic-antigravity' 'topic-antigravity-skills' 'topic-claude-code' 'topic-claude-code-skills' 'topic-codex-cli' 'topic-codex-skills' 'trace':1665 'track':999,1003,1004,1017,1046,1049,1056,1057,1746 'track.kind':1008 'transcrib':557,704,1208,1253 'transcriber.send':1121 'transcript':46,336,431,618,619,621,623,627,671,697,715,717,720,722,726,732,737,741,1209,1247 'treat':1876 'trigger':1639 'true':631,753,756,766,975,1168,1172,1270 'try/except':1634 'tts':666,803,848,866,1061,1157,1300,1312,1337,1419,1430,1433,1465,1472 'tts.convert':1482 'tts.flush':887 'tts.send':876 'tts.synthesize':1479 'tune':143,1535 'turbo':818,862 'turn':340 'twilio':225,1765,1777 'twiliophonenumb':644 'type':312,342,359,371,375,391,410,424,434,443,463,466,578,609,656,1147 'url':282,301,659,928 'use':120,256,489,676,903,1135,1395,1425,1441,1454,1478,1564,1712,1801,1861 'user':96,438,735,1250,1260,1262,1278,1459,1538,1582,1802,1808,1814,1821,1828,1834,1839,1844,1849,1854 'user_text.strip':1256 'utter':760 'v':710 'v1':297 'v2':819,863 'vad':344,764,1165,1397,1400,1528,1533 'valid':1460,1886 'vapi':40,148,219,478,485,496,510,572,1770,1838 'vapi.vapi':515 'vari':197 'voic':2,8,15,16,22,29,33,42,69,72,84,101,134,149,154,164,239,247,249,261,280,318,345,479,482,499,543,679,767,811,856,899,909,991,1052,1124,1128,1138,1392,1578,1650,1663,1671,1685,1688,1692,1708,1721,1724,1728,1735,1762,1806,1812,1833 'voice-ai-develop':1,1691,1727,1761 'voice-en':32 'voice-to-voic':246 'voiceag':1174 'voiceagentconfig':1152,1179 'voiceid':546 'want':259 'weather':363,366 'web':233,648,650,657,1675,1682 'webhook':502,563,573 'webrtc':57,161,228,661,896,1396 'websocket':174,229,269,842,849,1450,1508,1514,1618,1625 'websockets.connect':300 'well':1792 'whisper':338 'without':263,1509,1597,1619 'work':1791 'workflow':1642,1699,1732,1766 'wrap':1632 'ws':306,404 'ws.send':310,389,461 'yield':838,879,888,1244,1359,1376 'your-livekit.livekit.cloud':929,1032","prices":[{"id":"58ab85a2-fa1b-41a7-82c5-6f0864a3b627","listingId":"0d8c322d-0bbf-4077-8a98-79b4c09fcab0","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"sickn33","category":"antigravity-awesome-skills","install_from":"skills.sh"},"createdAt":"2026-04-18T20:38:59.220Z"}],"sources":[{"listingId":"0d8c322d-0bbf-4077-8a98-79b4c09fcab0","source":"github","sourceId":"sickn33/antigravity-awesome-skills/voice-ai-development","sourceUrl":"https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/voice-ai-development","isPrimary":false,"firstSeenAt":"2026-04-18T21:47:18.565Z","lastSeenAt":"2026-04-25T06:52:16.742Z"},{"listingId":"0d8c322d-0bbf-4077-8a98-79b4c09fcab0","source":"skills_sh","sourceId":"sickn33/antigravity-awesome-skills/voice-ai-development","sourceUrl":"https://skills.sh/sickn33/antigravity-awesome-skills/voice-ai-development","isPrimary":true,"firstSeenAt":"2026-04-18T20:38:59.220Z","lastSeenAt":"2026-04-23T00:40:55.296Z"}],"details":{"listingId":"0d8c322d-0bbf-4077-8a98-79b4c09fcab0","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"sickn33","slug":"voice-ai-development","github":{"repo":"sickn33/antigravity-awesome-skills","stars":34997,"topics":["agent-skills","agentic-skills","ai-agent-skills","ai-agents","ai-coding","ai-workflows","antigravity","antigravity-skills","claude-code","claude-code-skills","codex-cli","codex-skills","cursor","cursor-skills","developer-tools","gemini-cli","gemini-skills","kiro","mcp","skill-library"],"license":"mit","html_url":"https://github.com/sickn33/antigravity-awesome-skills","pushed_at":"2026-04-25T06:33:17Z","description":"Installable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.","skill_md_sha":"7295977cb8d65be3e31c4988838a5a709d809398","skill_md_path":"skills/voice-ai-development/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/voice-ai-development"},"layout":"multi","source":"github","category":"antigravity-awesome-skills","frontmatter":{"name":"voice-ai-development","description":"Expert in building voice AI applications - from real-time voice"},"skills_sh_url":"https://skills.sh/sickn33/antigravity-awesome-skills/voice-ai-development"},"updatedAt":"2026-04-25T06:52:16.742Z"}}