[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-openai-twilio-voice-conversation-relay":3,"mdc--p497rh-key":39,"related-repo-openai-twilio-voice-conversation-relay":1600,"related-org-openai-twilio-voice-conversation-relay":1719},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":28,"repoUrl":29,"updatedAt":30,"license":31,"forks":32,"topics":33,"repo":34,"sourceUrl":37,"mdContent":38},"twilio-voice-conversation-relay","build voice agents with Twilio","Build AI-powered voice agents using Twilio ConversationRelay. Handles real-time speech recognition (ASR), text-to-speech (TTS), and bidirectional audio streaming via WebSocket. Covers TwiML setup, WebSocket message types, LLM integration, streaming responses, and voice provider configuration. Use this skill to build voice bots, IVR replacements, or real-time AI voice assistants on Twilio calls.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"openai","OpenAI","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fopenai.png",[12,16,19,22,25],{"name":13,"slug":14,"type":15},"WebSockets","websockets","tag",{"name":17,"slug":18,"type":15},"Agents","agents",{"name":20,"slug":21,"type":15},"Audio","audio",{"name":23,"slug":24,"type":15},"Speech","speech",{"name":26,"slug":27,"type":15},"Twilio","twilio",3992,"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fplugins","2026-06-30T19:00:57.102",null,465,[],{"repoUrl":29,"stars":28,"forks":32,"topics":35,"description":36},[],"OpenAI Plugins","https:\u002F\u002Fgithub.com\u002Fopenai\u002Fplugins\u002Ftree\u002FHEAD\u002Fplugins\u002Ftwilio-developer-kit\u002Fskills\u002Ftwilio-voice-conversation-relay","---\nname: twilio-voice-conversation-relay\ndescription: >\n  Build AI-powered voice agents using Twilio ConversationRelay. Handles\n  real-time speech recognition (ASR), text-to-speech (TTS), and bidirectional\n  audio streaming via WebSocket. Covers TwiML setup, WebSocket message types,\n  LLM integration, streaming responses, and voice provider configuration. Use\n  this skill to build voice bots, IVR replacements, or real-time AI voice\n  assistants on Twilio calls.\n---\n\n## Overview\n\nConversationRelay connects Twilio's telephony layer to your app via a persistent WebSocket. Twilio handles ASR (speech-to-text) and TTS (text-to-speech); your app receives transcripts, calls an LLM, and sends text back for playback.\n\n```\nCaller ←→ Twilio (ASR\u002FTTS) ←→ WebSocket ←→ Your App ←→ LLM\n```\n\n---\n\n## Prerequisites\n\n- Upgraded Twilio account with ConversationRelay access (requires onboarding)\n  — New to Twilio? See `twilio-account-setup`\n  — Start onboarding at: [Console > Voice > ConversationRelay](https:\u002F\u002Fconsole.twilio.com\u002Fus1\u002Fvoice\u002Fconversation-relay) — access is **not** instant\n- A voice-capable Twilio phone number\n- `TWILIO_ACCOUNT_SID` and `TWILIO_AUTH_TOKEN` — see `twilio-iam-auth-setup`\n- WebSocket server reachable via `wss:\u002F\u002F` (TLS required)\n- An LLM integration (OpenAI, Anthropic, etc.)\n- For placing calls: see `twilio-voice-outbound-calls`\n\n**Onboarding:** Complete via Console > Voice > ConversationRelay > Onboarding. Select TTS\u002FASR providers:\n- **TTS:** Deepgram, Amazon Polly, Google Cloud TTS, ElevenLabs\n- **ASR:** Deepgram, Google Cloud STT\n\n---\n\n## Quickstart\n\n**Step 1 — Return TwiML pointing to your WebSocket server**\n\n**Python (Flask)**\n```python\nfrom flask import Flask\nfrom twilio.twiml.voice_response import VoiceResponse, Connect, ConversationRelay\n\napp = Flask(__name__)\n\n@app.route(\"\u002Fvoice\", methods=[\"POST\"])\ndef voice():\n    response = VoiceResponse()\n    connect = Connect()\n    connect.conversation_relay(\n        url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n        welcome_greeting=\"Hello! How can I help you today?\"\n    )\n    response.append(connect)\n    return str(response)\n```\n\n**Node.js (Express)**\n```node\nconst { VoiceResponse } = require(\"twilio\").twiml;\n\napp.post(\"\u002Fvoice\", (req, res) => {\n    const response = new VoiceResponse();\n    const connect = response.connect();\n    connect.conversationRelay({\n        url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n        welcomeGreeting: \"Hello! How can I help you today?\",\n    });\n    res.type(\"text\u002Fxml\").send(response.toString());\n});\n```\n\n**Step 2 — Handle WebSocket events and respond with text**\n\n**Python (websockets)**\n```python\nimport asyncio, json, websockets\n\nasync def handle_call(websocket):\n    async for message in websocket:\n        event = json.loads(message)\n        if event[\"type\"] == \"prompt\":\n            ai_response = await call_llm(event[\"voicePrompt\"])\n            await websocket.send(json.dumps({\"type\": \"text\", \"token\": ai_response, \"last\": True}))\n\nasync def main():\n    async with websockets.serve(handle_call, \"0.0.0.0\", 8080):\n        await asyncio.Future()\n\nasyncio.run(main())\n```\n\n**Node.js (ws)**\n```node\nconst WebSocket = require(\"ws\");\nconst wss = new WebSocket.Server({ port: 8080 });\n\nwss.on(\"connection\", (ws) => {\n    ws.on(\"message\", async (data) => {\n        const event = JSON.parse(data);\n        if (event.type === \"prompt\") {\n            const aiResponse = await callLLM(event.voicePrompt);\n            ws.send(JSON.stringify({ type: \"text\", token: aiResponse, last: true }));\n        }\n    });\n});\n```\n\n> **Security:** The `voicePrompt` field contains ASR-transcribed caller speech — it is untrusted external input. When passing to an LLM, isolate it as user input within a structured system prompt. Implement topic boundaries and output filtering to prevent the LLM from disclosing system instructions or speaking inappropriate content. ConversationRelay is a pure transport layer with no built-in content safety — any LLM output is spoken to the caller verbatim.\n\n---\n\n## Key Patterns\n\n### WebSocket Message Types\n\n**Received from Twilio:**\n\n| Type | When | Key fields |\n|------|------|-----------|\n| `connected` | WebSocket opened | `callSid`, `streamSid` |\n| `prompt` | User finished speaking | `voicePrompt` (transcript) |\n| `interrupt` | User interrupted TTS | — |\n| `dtmf` | User pressed keypad key | `digit` |\n| `error` | An error occurred | `description` |\n\n**Sent to Twilio:**\n\n| Type | Purpose | Key fields |\n|------|---------|-----------|\n| `text` | Send TTS response | `token` (text), `last` (bool) |\n| `interrupt` | Stop current TTS | — |\n| `end` | Hang up the call | `reason` |\n\n### Stream LLM Responses Token-by-Token\n\nLower latency by streaming as the LLM generates output — Twilio starts speaking before the full response is ready.\n\n**Python**\n```python\nasync for chunk in llm_stream:\n    await websocket.send(json.dumps({\"type\": \"text\", \"token\": chunk, \"last\": False}))\nawait websocket.send(json.dumps({\"type\": \"text\", \"token\": \"\", \"last\": True}))\n```\n\n**Node.js**\n```node\nfor await (const chunk of llmStream) {\n    ws.send(JSON.stringify({ type: \"text\", token: chunk, last: false }));\n}\nws.send(JSON.stringify({ type: \"text\", token: \"\", last: true }));\n```\n\n### Voice Configuration\n\n**Python**\n```python\nconnect.conversation_relay(\n    url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n    voice=\"en-US-Neural2-F\",\n    language=\"en-US\",\n    transcription_provider=\"deepgram\",\n    speech_model=\"nova-2-phonecall\",\n    interrupt_by_dtmf=True,\n)\n```\n\n**Node.js**\n```node\nconnect.conversationRelay({\n    url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n    voice: \"en-US-Neural2-F\",\n    language: \"en-US\",\n    transcriptionProvider: \"deepgram\",\n    speechModel: \"nova-2-phonecall\",\n    interruptByDtmf: true,\n});\n```\n\n---\n\n## CANNOT\n\n- **No raw audio access** — Text in, text out only. For raw audio, use `\u003CConnect>\u003CStream>` (Media Streams).\n- **Cannot mix with Media Streams** — `\u003CConnect>\u003CStream>` and `\u003CConnect>\u003CConversationRelay>` are mutually exclusive on the same call. No error — one is silently ignored.\n- **No custom STT\u002FTTS engines** — Limited to Deepgram + Google (STT) and Deepgram + Amazon Polly + Google + ElevenLabs (TTS).\n- **No mid-session voice\u002Fprovider changes** — Voice and provider are set at TwiML time. Only `language` can be switched mid-session via WebSocket message.\n- **No WebSocket auto-reconnection** — If the WebSocket drops, the call disconnects. Implement recovery via `\u003CConnect action>` URL.\n- **Voice only** — No SMS\u002Fmessaging support. For omnichannel, use Conversation Orchestrator.\n- **No built-in memory or context** — BYO conversation history and context management.\n- **No LLM integration** — Pure transport layer. You bring your own LLM via the WebSocket server.\n- **No server-side recording via REST API** — `record:true` on the Calls API is silently ignored. Must use `\u003CStart>\u003CRecording>` before `\u003CConnect>` in TwiML.\n- **Not PCI compliant with Voice Intelligence v2** — Do not enable `intelligenceService` in PCI workflows.\n- **ElevenLabs requires account enablement** — Accounts without ElevenLabs access get error 64101. Voice IDs (not human-readable names) are required.\n- **Cannot use ConversationRelay without onboarding** — Not available immediately on a new account\n- **Cannot use non-TLS WebSocket** — Server must be reachable via `wss:\u002F\u002F` (TLS required)\n- **Cannot stream audio without `last: true`** — Twilio won't play audio until it sees a `last: true` token in the stream\n\n---\n\n## Next Steps\n\n- **Place outbound calls:** `twilio-voice-outbound-calls`\n- **Standard IVR without AI:** `twilio-voice-twiml`\n",{"data":40,"body":41},{"name":4,"description":6},{"type":42,"children":43},"root",[44,53,59,72,76,82,178,188,211,214,220,228,236,382,390,486,494,502,618,626,726,748,751,757,764,772,940,948,1054,1060,1065,1073,1104,1112,1151,1157,1164,1235,1242,1312,1315,1321,1552,1555,1561,1594],{"type":45,"tag":46,"props":47,"children":49},"element","h2",{"id":48},"overview",[50],{"type":51,"value":52},"text","Overview",{"type":45,"tag":54,"props":55,"children":56},"p",{},[57],{"type":51,"value":58},"ConversationRelay connects Twilio's telephony layer to your app via a persistent WebSocket. Twilio handles ASR (speech-to-text) and TTS (text-to-speech); your app receives transcripts, calls an LLM, and sends text back for playback.",{"type":45,"tag":60,"props":61,"children":65},"pre",{"className":62,"code":64,"language":51},[63],"language-text","Caller ←→ Twilio (ASR\u002FTTS) ←→ WebSocket ←→ Your App ←→ LLM\n",[66],{"type":45,"tag":67,"props":68,"children":70},"code",{"__ignoreMap":69},"",[71],{"type":51,"value":64},{"type":45,"tag":73,"props":74,"children":75},"hr",{},[],{"type":45,"tag":46,"props":77,"children":79},{"id":78},"prerequisites",[80],{"type":51,"value":81},"Prerequisites",{"type":45,"tag":83,"props":84,"children":85},"ul",{},[86,119,124,149,162,167],{"type":45,"tag":87,"props":88,"children":89},"li",{},[90,92,98,100,109,111,117],{"type":51,"value":91},"Upgraded Twilio account with ConversationRelay access (requires onboarding)\n— New to Twilio? See ",{"type":45,"tag":67,"props":93,"children":95},{"className":94},[],[96],{"type":51,"value":97},"twilio-account-setup",{"type":51,"value":99},"\n— Start onboarding at: ",{"type":45,"tag":101,"props":102,"children":106},"a",{"href":103,"rel":104},"https:\u002F\u002Fconsole.twilio.com\u002Fus1\u002Fvoice\u002Fconversation-relay",[105],"nofollow",[107],{"type":51,"value":108},"Console > Voice > ConversationRelay",{"type":51,"value":110}," — access is ",{"type":45,"tag":112,"props":113,"children":114},"strong",{},[115],{"type":51,"value":116},"not",{"type":51,"value":118}," instant",{"type":45,"tag":87,"props":120,"children":121},{},[122],{"type":51,"value":123},"A voice-capable Twilio phone number",{"type":45,"tag":87,"props":125,"children":126},{},[127,133,135,141,143],{"type":45,"tag":67,"props":128,"children":130},{"className":129},[],[131],{"type":51,"value":132},"TWILIO_ACCOUNT_SID",{"type":51,"value":134}," and ",{"type":45,"tag":67,"props":136,"children":138},{"className":137},[],[139],{"type":51,"value":140},"TWILIO_AUTH_TOKEN",{"type":51,"value":142}," — see ",{"type":45,"tag":67,"props":144,"children":146},{"className":145},[],[147],{"type":51,"value":148},"twilio-iam-auth-setup",{"type":45,"tag":87,"props":150,"children":151},{},[152,154,160],{"type":51,"value":153},"WebSocket server reachable via ",{"type":45,"tag":67,"props":155,"children":157},{"className":156},[],[158],{"type":51,"value":159},"wss:\u002F\u002F",{"type":51,"value":161}," (TLS required)",{"type":45,"tag":87,"props":163,"children":164},{},[165],{"type":51,"value":166},"An LLM integration (OpenAI, Anthropic, etc.)",{"type":45,"tag":87,"props":168,"children":169},{},[170,172],{"type":51,"value":171},"For placing calls: see ",{"type":45,"tag":67,"props":173,"children":175},{"className":174},[],[176],{"type":51,"value":177},"twilio-voice-outbound-calls",{"type":45,"tag":54,"props":179,"children":180},{},[181,186],{"type":45,"tag":112,"props":182,"children":183},{},[184],{"type":51,"value":185},"Onboarding:",{"type":51,"value":187}," Complete via Console > Voice > ConversationRelay > Onboarding. Select TTS\u002FASR providers:",{"type":45,"tag":83,"props":189,"children":190},{},[191,201],{"type":45,"tag":87,"props":192,"children":193},{},[194,199],{"type":45,"tag":112,"props":195,"children":196},{},[197],{"type":51,"value":198},"TTS:",{"type":51,"value":200}," Deepgram, Amazon Polly, Google Cloud TTS, ElevenLabs",{"type":45,"tag":87,"props":202,"children":203},{},[204,209],{"type":45,"tag":112,"props":205,"children":206},{},[207],{"type":51,"value":208},"ASR:",{"type":51,"value":210}," Deepgram, Google Cloud STT",{"type":45,"tag":73,"props":212,"children":213},{},[],{"type":45,"tag":46,"props":215,"children":217},{"id":216},"quickstart",[218],{"type":51,"value":219},"Quickstart",{"type":45,"tag":54,"props":221,"children":222},{},[223],{"type":45,"tag":112,"props":224,"children":225},{},[226],{"type":51,"value":227},"Step 1 — Return TwiML pointing to your WebSocket server",{"type":45,"tag":54,"props":229,"children":230},{},[231],{"type":45,"tag":112,"props":232,"children":233},{},[234],{"type":51,"value":235},"Python (Flask)",{"type":45,"tag":60,"props":237,"children":241},{"className":238,"code":239,"language":240,"meta":69,"style":69},"language-python shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","from flask import Flask\nfrom twilio.twiml.voice_response import VoiceResponse, Connect, ConversationRelay\n\napp = Flask(__name__)\n\n@app.route(\"\u002Fvoice\", methods=[\"POST\"])\ndef voice():\n    response = VoiceResponse()\n    connect = Connect()\n    connect.conversation_relay(\n        url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n        welcome_greeting=\"Hello! How can I help you today?\"\n    )\n    response.append(connect)\n    return str(response)\n","python",[242],{"type":45,"tag":67,"props":243,"children":244},{"__ignoreMap":69},[245,256,265,275,284,292,301,310,319,328,337,346,355,364,373],{"type":45,"tag":246,"props":247,"children":250},"span",{"class":248,"line":249},"line",1,[251],{"type":45,"tag":246,"props":252,"children":253},{},[254],{"type":51,"value":255},"from flask import Flask\n",{"type":45,"tag":246,"props":257,"children":259},{"class":248,"line":258},2,[260],{"type":45,"tag":246,"props":261,"children":262},{},[263],{"type":51,"value":264},"from twilio.twiml.voice_response import VoiceResponse, Connect, ConversationRelay\n",{"type":45,"tag":246,"props":266,"children":268},{"class":248,"line":267},3,[269],{"type":45,"tag":246,"props":270,"children":272},{"emptyLinePlaceholder":271},true,[273],{"type":51,"value":274},"\n",{"type":45,"tag":246,"props":276,"children":278},{"class":248,"line":277},4,[279],{"type":45,"tag":246,"props":280,"children":281},{},[282],{"type":51,"value":283},"app = Flask(__name__)\n",{"type":45,"tag":246,"props":285,"children":287},{"class":248,"line":286},5,[288],{"type":45,"tag":246,"props":289,"children":290},{"emptyLinePlaceholder":271},[291],{"type":51,"value":274},{"type":45,"tag":246,"props":293,"children":295},{"class":248,"line":294},6,[296],{"type":45,"tag":246,"props":297,"children":298},{},[299],{"type":51,"value":300},"@app.route(\"\u002Fvoice\", methods=[\"POST\"])\n",{"type":45,"tag":246,"props":302,"children":304},{"class":248,"line":303},7,[305],{"type":45,"tag":246,"props":306,"children":307},{},[308],{"type":51,"value":309},"def voice():\n",{"type":45,"tag":246,"props":311,"children":313},{"class":248,"line":312},8,[314],{"type":45,"tag":246,"props":315,"children":316},{},[317],{"type":51,"value":318},"    response = VoiceResponse()\n",{"type":45,"tag":246,"props":320,"children":322},{"class":248,"line":321},9,[323],{"type":45,"tag":246,"props":324,"children":325},{},[326],{"type":51,"value":327},"    connect = Connect()\n",{"type":45,"tag":246,"props":329,"children":331},{"class":248,"line":330},10,[332],{"type":45,"tag":246,"props":333,"children":334},{},[335],{"type":51,"value":336},"    connect.conversation_relay(\n",{"type":45,"tag":246,"props":338,"children":340},{"class":248,"line":339},11,[341],{"type":45,"tag":246,"props":342,"children":343},{},[344],{"type":51,"value":345},"        url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n",{"type":45,"tag":246,"props":347,"children":349},{"class":248,"line":348},12,[350],{"type":45,"tag":246,"props":351,"children":352},{},[353],{"type":51,"value":354},"        welcome_greeting=\"Hello! How can I help you today?\"\n",{"type":45,"tag":246,"props":356,"children":358},{"class":248,"line":357},13,[359],{"type":45,"tag":246,"props":360,"children":361},{},[362],{"type":51,"value":363},"    )\n",{"type":45,"tag":246,"props":365,"children":367},{"class":248,"line":366},14,[368],{"type":45,"tag":246,"props":369,"children":370},{},[371],{"type":51,"value":372},"    response.append(connect)\n",{"type":45,"tag":246,"props":374,"children":376},{"class":248,"line":375},15,[377],{"type":45,"tag":246,"props":378,"children":379},{},[380],{"type":51,"value":381},"    return str(response)\n",{"type":45,"tag":54,"props":383,"children":384},{},[385],{"type":45,"tag":112,"props":386,"children":387},{},[388],{"type":51,"value":389},"Node.js (Express)",{"type":45,"tag":60,"props":391,"children":395},{"className":392,"code":393,"language":394,"meta":69,"style":69},"language-node shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","const { VoiceResponse } = require(\"twilio\").twiml;\n\napp.post(\"\u002Fvoice\", (req, res) => {\n    const response = new VoiceResponse();\n    const connect = response.connect();\n    connect.conversationRelay({\n        url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n        welcomeGreeting: \"Hello! How can I help you today?\",\n    });\n    res.type(\"text\u002Fxml\").send(response.toString());\n});\n","node",[396],{"type":45,"tag":67,"props":397,"children":398},{"__ignoreMap":69},[399,407,414,422,430,438,446,454,462,470,478],{"type":45,"tag":246,"props":400,"children":401},{"class":248,"line":249},[402],{"type":45,"tag":246,"props":403,"children":404},{},[405],{"type":51,"value":406},"const { VoiceResponse } = require(\"twilio\").twiml;\n",{"type":45,"tag":246,"props":408,"children":409},{"class":248,"line":258},[410],{"type":45,"tag":246,"props":411,"children":412},{"emptyLinePlaceholder":271},[413],{"type":51,"value":274},{"type":45,"tag":246,"props":415,"children":416},{"class":248,"line":267},[417],{"type":45,"tag":246,"props":418,"children":419},{},[420],{"type":51,"value":421},"app.post(\"\u002Fvoice\", (req, res) => {\n",{"type":45,"tag":246,"props":423,"children":424},{"class":248,"line":277},[425],{"type":45,"tag":246,"props":426,"children":427},{},[428],{"type":51,"value":429},"    const response = new VoiceResponse();\n",{"type":45,"tag":246,"props":431,"children":432},{"class":248,"line":286},[433],{"type":45,"tag":246,"props":434,"children":435},{},[436],{"type":51,"value":437},"    const connect = response.connect();\n",{"type":45,"tag":246,"props":439,"children":440},{"class":248,"line":294},[441],{"type":45,"tag":246,"props":442,"children":443},{},[444],{"type":51,"value":445},"    connect.conversationRelay({\n",{"type":45,"tag":246,"props":447,"children":448},{"class":248,"line":303},[449],{"type":45,"tag":246,"props":450,"children":451},{},[452],{"type":51,"value":453},"        url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n",{"type":45,"tag":246,"props":455,"children":456},{"class":248,"line":312},[457],{"type":45,"tag":246,"props":458,"children":459},{},[460],{"type":51,"value":461},"        welcomeGreeting: \"Hello! How can I help you today?\",\n",{"type":45,"tag":246,"props":463,"children":464},{"class":248,"line":321},[465],{"type":45,"tag":246,"props":466,"children":467},{},[468],{"type":51,"value":469},"    });\n",{"type":45,"tag":246,"props":471,"children":472},{"class":248,"line":330},[473],{"type":45,"tag":246,"props":474,"children":475},{},[476],{"type":51,"value":477},"    res.type(\"text\u002Fxml\").send(response.toString());\n",{"type":45,"tag":246,"props":479,"children":480},{"class":248,"line":339},[481],{"type":45,"tag":246,"props":482,"children":483},{},[484],{"type":51,"value":485},"});\n",{"type":45,"tag":54,"props":487,"children":488},{},[489],{"type":45,"tag":112,"props":490,"children":491},{},[492],{"type":51,"value":493},"Step 2 — Handle WebSocket events and respond with text",{"type":45,"tag":54,"props":495,"children":496},{},[497],{"type":45,"tag":112,"props":498,"children":499},{},[500],{"type":51,"value":501},"Python (websockets)",{"type":45,"tag":60,"props":503,"children":505},{"className":238,"code":504,"language":240,"meta":69,"style":69},"import asyncio, json, websockets\n\nasync def handle_call(websocket):\n    async for message in websocket:\n        event = json.loads(message)\n        if event[\"type\"] == \"prompt\":\n            ai_response = await call_llm(event[\"voicePrompt\"])\n            await websocket.send(json.dumps({\"type\": \"text\", \"token\": ai_response, \"last\": True}))\n\nasync def main():\n    async with websockets.serve(handle_call, \"0.0.0.0\", 8080):\n        await asyncio.Future()\n\nasyncio.run(main())\n",[506],{"type":45,"tag":67,"props":507,"children":508},{"__ignoreMap":69},[509,517,524,532,540,548,556,564,572,579,587,595,603,610],{"type":45,"tag":246,"props":510,"children":511},{"class":248,"line":249},[512],{"type":45,"tag":246,"props":513,"children":514},{},[515],{"type":51,"value":516},"import asyncio, json, websockets\n",{"type":45,"tag":246,"props":518,"children":519},{"class":248,"line":258},[520],{"type":45,"tag":246,"props":521,"children":522},{"emptyLinePlaceholder":271},[523],{"type":51,"value":274},{"type":45,"tag":246,"props":525,"children":526},{"class":248,"line":267},[527],{"type":45,"tag":246,"props":528,"children":529},{},[530],{"type":51,"value":531},"async def handle_call(websocket):\n",{"type":45,"tag":246,"props":533,"children":534},{"class":248,"line":277},[535],{"type":45,"tag":246,"props":536,"children":537},{},[538],{"type":51,"value":539},"    async for message in websocket:\n",{"type":45,"tag":246,"props":541,"children":542},{"class":248,"line":286},[543],{"type":45,"tag":246,"props":544,"children":545},{},[546],{"type":51,"value":547},"        event = json.loads(message)\n",{"type":45,"tag":246,"props":549,"children":550},{"class":248,"line":294},[551],{"type":45,"tag":246,"props":552,"children":553},{},[554],{"type":51,"value":555},"        if event[\"type\"] == \"prompt\":\n",{"type":45,"tag":246,"props":557,"children":558},{"class":248,"line":303},[559],{"type":45,"tag":246,"props":560,"children":561},{},[562],{"type":51,"value":563},"            ai_response = await call_llm(event[\"voicePrompt\"])\n",{"type":45,"tag":246,"props":565,"children":566},{"class":248,"line":312},[567],{"type":45,"tag":246,"props":568,"children":569},{},[570],{"type":51,"value":571},"            await websocket.send(json.dumps({\"type\": \"text\", \"token\": ai_response, \"last\": True}))\n",{"type":45,"tag":246,"props":573,"children":574},{"class":248,"line":321},[575],{"type":45,"tag":246,"props":576,"children":577},{"emptyLinePlaceholder":271},[578],{"type":51,"value":274},{"type":45,"tag":246,"props":580,"children":581},{"class":248,"line":330},[582],{"type":45,"tag":246,"props":583,"children":584},{},[585],{"type":51,"value":586},"async def main():\n",{"type":45,"tag":246,"props":588,"children":589},{"class":248,"line":339},[590],{"type":45,"tag":246,"props":591,"children":592},{},[593],{"type":51,"value":594},"    async with websockets.serve(handle_call, \"0.0.0.0\", 8080):\n",{"type":45,"tag":246,"props":596,"children":597},{"class":248,"line":348},[598],{"type":45,"tag":246,"props":599,"children":600},{},[601],{"type":51,"value":602},"        await asyncio.Future()\n",{"type":45,"tag":246,"props":604,"children":605},{"class":248,"line":357},[606],{"type":45,"tag":246,"props":607,"children":608},{"emptyLinePlaceholder":271},[609],{"type":51,"value":274},{"type":45,"tag":246,"props":611,"children":612},{"class":248,"line":366},[613],{"type":45,"tag":246,"props":614,"children":615},{},[616],{"type":51,"value":617},"asyncio.run(main())\n",{"type":45,"tag":54,"props":619,"children":620},{},[621],{"type":45,"tag":112,"props":622,"children":623},{},[624],{"type":51,"value":625},"Node.js (ws)",{"type":45,"tag":60,"props":627,"children":629},{"className":392,"code":628,"language":394,"meta":69,"style":69},"const WebSocket = require(\"ws\");\nconst wss = new WebSocket.Server({ port: 8080 });\n\nwss.on(\"connection\", (ws) => {\n    ws.on(\"message\", async (data) => {\n        const event = JSON.parse(data);\n        if (event.type === \"prompt\") {\n            const aiResponse = await callLLM(event.voicePrompt);\n            ws.send(JSON.stringify({ type: \"text\", token: aiResponse, last: true }));\n        }\n    });\n});\n",[630],{"type":45,"tag":67,"props":631,"children":632},{"__ignoreMap":69},[633,641,649,656,664,672,680,688,696,704,712,719],{"type":45,"tag":246,"props":634,"children":635},{"class":248,"line":249},[636],{"type":45,"tag":246,"props":637,"children":638},{},[639],{"type":51,"value":640},"const WebSocket = require(\"ws\");\n",{"type":45,"tag":246,"props":642,"children":643},{"class":248,"line":258},[644],{"type":45,"tag":246,"props":645,"children":646},{},[647],{"type":51,"value":648},"const wss = new WebSocket.Server({ port: 8080 });\n",{"type":45,"tag":246,"props":650,"children":651},{"class":248,"line":267},[652],{"type":45,"tag":246,"props":653,"children":654},{"emptyLinePlaceholder":271},[655],{"type":51,"value":274},{"type":45,"tag":246,"props":657,"children":658},{"class":248,"line":277},[659],{"type":45,"tag":246,"props":660,"children":661},{},[662],{"type":51,"value":663},"wss.on(\"connection\", (ws) => {\n",{"type":45,"tag":246,"props":665,"children":666},{"class":248,"line":286},[667],{"type":45,"tag":246,"props":668,"children":669},{},[670],{"type":51,"value":671},"    ws.on(\"message\", async (data) => {\n",{"type":45,"tag":246,"props":673,"children":674},{"class":248,"line":294},[675],{"type":45,"tag":246,"props":676,"children":677},{},[678],{"type":51,"value":679},"        const event = JSON.parse(data);\n",{"type":45,"tag":246,"props":681,"children":682},{"class":248,"line":303},[683],{"type":45,"tag":246,"props":684,"children":685},{},[686],{"type":51,"value":687},"        if (event.type === \"prompt\") {\n",{"type":45,"tag":246,"props":689,"children":690},{"class":248,"line":312},[691],{"type":45,"tag":246,"props":692,"children":693},{},[694],{"type":51,"value":695},"            const aiResponse = await callLLM(event.voicePrompt);\n",{"type":45,"tag":246,"props":697,"children":698},{"class":248,"line":321},[699],{"type":45,"tag":246,"props":700,"children":701},{},[702],{"type":51,"value":703},"            ws.send(JSON.stringify({ type: \"text\", token: aiResponse, last: true }));\n",{"type":45,"tag":246,"props":705,"children":706},{"class":248,"line":330},[707],{"type":45,"tag":246,"props":708,"children":709},{},[710],{"type":51,"value":711},"        }\n",{"type":45,"tag":246,"props":713,"children":714},{"class":248,"line":339},[715],{"type":45,"tag":246,"props":716,"children":717},{},[718],{"type":51,"value":469},{"type":45,"tag":246,"props":720,"children":721},{"class":248,"line":348},[722],{"type":45,"tag":246,"props":723,"children":724},{},[725],{"type":51,"value":485},{"type":45,"tag":727,"props":728,"children":729},"blockquote",{},[730],{"type":45,"tag":54,"props":731,"children":732},{},[733,738,740,746],{"type":45,"tag":112,"props":734,"children":735},{},[736],{"type":51,"value":737},"Security:",{"type":51,"value":739}," The ",{"type":45,"tag":67,"props":741,"children":743},{"className":742},[],[744],{"type":51,"value":745},"voicePrompt",{"type":51,"value":747}," field contains ASR-transcribed caller speech — it is untrusted external input. When passing to an LLM, isolate it as user input within a structured system prompt. Implement topic boundaries and output filtering to prevent the LLM from disclosing system instructions or speaking inappropriate content. ConversationRelay is a pure transport layer with no built-in content safety — any LLM output is spoken to the caller verbatim.",{"type":45,"tag":73,"props":749,"children":750},{},[],{"type":45,"tag":46,"props":752,"children":754},{"id":753},"key-patterns",[755],{"type":51,"value":756},"Key Patterns",{"type":45,"tag":758,"props":759,"children":761},"h3",{"id":760},"websocket-message-types",[762],{"type":51,"value":763},"WebSocket Message Types",{"type":45,"tag":54,"props":765,"children":766},{},[767],{"type":45,"tag":112,"props":768,"children":769},{},[770],{"type":51,"value":771},"Received from Twilio:",{"type":45,"tag":773,"props":774,"children":775},"table",{},[776,800],{"type":45,"tag":777,"props":778,"children":779},"thead",{},[780],{"type":45,"tag":781,"props":782,"children":783},"tr",{},[784,790,795],{"type":45,"tag":785,"props":786,"children":787},"th",{},[788],{"type":51,"value":789},"Type",{"type":45,"tag":785,"props":791,"children":792},{},[793],{"type":51,"value":794},"When",{"type":45,"tag":785,"props":796,"children":797},{},[798],{"type":51,"value":799},"Key fields",{"type":45,"tag":801,"props":802,"children":803},"tbody",{},[804,839,866,888,914],{"type":45,"tag":781,"props":805,"children":806},{},[807,817,822],{"type":45,"tag":808,"props":809,"children":810},"td",{},[811],{"type":45,"tag":67,"props":812,"children":814},{"className":813},[],[815],{"type":51,"value":816},"connected",{"type":45,"tag":808,"props":818,"children":819},{},[820],{"type":51,"value":821},"WebSocket opened",{"type":45,"tag":808,"props":823,"children":824},{},[825,831,833],{"type":45,"tag":67,"props":826,"children":828},{"className":827},[],[829],{"type":51,"value":830},"callSid",{"type":51,"value":832},", ",{"type":45,"tag":67,"props":834,"children":836},{"className":835},[],[837],{"type":51,"value":838},"streamSid",{"type":45,"tag":781,"props":840,"children":841},{},[842,851,856],{"type":45,"tag":808,"props":843,"children":844},{},[845],{"type":45,"tag":67,"props":846,"children":848},{"className":847},[],[849],{"type":51,"value":850},"prompt",{"type":45,"tag":808,"props":852,"children":853},{},[854],{"type":51,"value":855},"User finished speaking",{"type":45,"tag":808,"props":857,"children":858},{},[859,864],{"type":45,"tag":67,"props":860,"children":862},{"className":861},[],[863],{"type":51,"value":745},{"type":51,"value":865}," (transcript)",{"type":45,"tag":781,"props":867,"children":868},{},[869,878,883],{"type":45,"tag":808,"props":870,"children":871},{},[872],{"type":45,"tag":67,"props":873,"children":875},{"className":874},[],[876],{"type":51,"value":877},"interrupt",{"type":45,"tag":808,"props":879,"children":880},{},[881],{"type":51,"value":882},"User interrupted TTS",{"type":45,"tag":808,"props":884,"children":885},{},[886],{"type":51,"value":887},"—",{"type":45,"tag":781,"props":889,"children":890},{},[891,900,905],{"type":45,"tag":808,"props":892,"children":893},{},[894],{"type":45,"tag":67,"props":895,"children":897},{"className":896},[],[898],{"type":51,"value":899},"dtmf",{"type":45,"tag":808,"props":901,"children":902},{},[903],{"type":51,"value":904},"User pressed keypad key",{"type":45,"tag":808,"props":906,"children":907},{},[908],{"type":45,"tag":67,"props":909,"children":911},{"className":910},[],[912],{"type":51,"value":913},"digit",{"type":45,"tag":781,"props":915,"children":916},{},[917,926,931],{"type":45,"tag":808,"props":918,"children":919},{},[920],{"type":45,"tag":67,"props":921,"children":923},{"className":922},[],[924],{"type":51,"value":925},"error",{"type":45,"tag":808,"props":927,"children":928},{},[929],{"type":51,"value":930},"An error occurred",{"type":45,"tag":808,"props":932,"children":933},{},[934],{"type":45,"tag":67,"props":935,"children":937},{"className":936},[],[938],{"type":51,"value":939},"description",{"type":45,"tag":54,"props":941,"children":942},{},[943],{"type":45,"tag":112,"props":944,"children":945},{},[946],{"type":51,"value":947},"Sent to Twilio:",{"type":45,"tag":773,"props":949,"children":950},{},[951,970],{"type":45,"tag":777,"props":952,"children":953},{},[954],{"type":45,"tag":781,"props":955,"children":956},{},[957,961,966],{"type":45,"tag":785,"props":958,"children":959},{},[960],{"type":51,"value":789},{"type":45,"tag":785,"props":962,"children":963},{},[964],{"type":51,"value":965},"Purpose",{"type":45,"tag":785,"props":967,"children":968},{},[969],{"type":51,"value":799},{"type":45,"tag":801,"props":971,"children":972},{},[973,1008,1028],{"type":45,"tag":781,"props":974,"children":975},{},[976,984,989],{"type":45,"tag":808,"props":977,"children":978},{},[979],{"type":45,"tag":67,"props":980,"children":982},{"className":981},[],[983],{"type":51,"value":51},{"type":45,"tag":808,"props":985,"children":986},{},[987],{"type":51,"value":988},"Send TTS response",{"type":45,"tag":808,"props":990,"children":991},{},[992,998,1000,1006],{"type":45,"tag":67,"props":993,"children":995},{"className":994},[],[996],{"type":51,"value":997},"token",{"type":51,"value":999}," (text), ",{"type":45,"tag":67,"props":1001,"children":1003},{"className":1002},[],[1004],{"type":51,"value":1005},"last",{"type":51,"value":1007}," (bool)",{"type":45,"tag":781,"props":1009,"children":1010},{},[1011,1019,1024],{"type":45,"tag":808,"props":1012,"children":1013},{},[1014],{"type":45,"tag":67,"props":1015,"children":1017},{"className":1016},[],[1018],{"type":51,"value":877},{"type":45,"tag":808,"props":1020,"children":1021},{},[1022],{"type":51,"value":1023},"Stop current TTS",{"type":45,"tag":808,"props":1025,"children":1026},{},[1027],{"type":51,"value":887},{"type":45,"tag":781,"props":1029,"children":1030},{},[1031,1040,1045],{"type":45,"tag":808,"props":1032,"children":1033},{},[1034],{"type":45,"tag":67,"props":1035,"children":1037},{"className":1036},[],[1038],{"type":51,"value":1039},"end",{"type":45,"tag":808,"props":1041,"children":1042},{},[1043],{"type":51,"value":1044},"Hang up the call",{"type":45,"tag":808,"props":1046,"children":1047},{},[1048],{"type":45,"tag":67,"props":1049,"children":1051},{"className":1050},[],[1052],{"type":51,"value":1053},"reason",{"type":45,"tag":758,"props":1055,"children":1057},{"id":1056},"stream-llm-responses-token-by-token",[1058],{"type":51,"value":1059},"Stream LLM Responses Token-by-Token",{"type":45,"tag":54,"props":1061,"children":1062},{},[1063],{"type":51,"value":1064},"Lower latency by streaming as the LLM generates output — Twilio starts speaking before the full response is ready.",{"type":45,"tag":54,"props":1066,"children":1067},{},[1068],{"type":45,"tag":112,"props":1069,"children":1070},{},[1071],{"type":51,"value":1072},"Python",{"type":45,"tag":60,"props":1074,"children":1076},{"className":238,"code":1075,"language":240,"meta":69,"style":69},"async for chunk in llm_stream:\n    await websocket.send(json.dumps({\"type\": \"text\", \"token\": chunk, \"last\": False}))\nawait websocket.send(json.dumps({\"type\": \"text\", \"token\": \"\", \"last\": True}))\n",[1077],{"type":45,"tag":67,"props":1078,"children":1079},{"__ignoreMap":69},[1080,1088,1096],{"type":45,"tag":246,"props":1081,"children":1082},{"class":248,"line":249},[1083],{"type":45,"tag":246,"props":1084,"children":1085},{},[1086],{"type":51,"value":1087},"async for chunk in llm_stream:\n",{"type":45,"tag":246,"props":1089,"children":1090},{"class":248,"line":258},[1091],{"type":45,"tag":246,"props":1092,"children":1093},{},[1094],{"type":51,"value":1095},"    await websocket.send(json.dumps({\"type\": \"text\", \"token\": chunk, \"last\": False}))\n",{"type":45,"tag":246,"props":1097,"children":1098},{"class":248,"line":267},[1099],{"type":45,"tag":246,"props":1100,"children":1101},{},[1102],{"type":51,"value":1103},"await websocket.send(json.dumps({\"type\": \"text\", \"token\": \"\", \"last\": True}))\n",{"type":45,"tag":54,"props":1105,"children":1106},{},[1107],{"type":45,"tag":112,"props":1108,"children":1109},{},[1110],{"type":51,"value":1111},"Node.js",{"type":45,"tag":60,"props":1113,"children":1115},{"className":392,"code":1114,"language":394,"meta":69,"style":69},"for await (const chunk of llmStream) {\n    ws.send(JSON.stringify({ type: \"text\", token: chunk, last: false }));\n}\nws.send(JSON.stringify({ type: \"text\", token: \"\", last: true }));\n",[1116],{"type":45,"tag":67,"props":1117,"children":1118},{"__ignoreMap":69},[1119,1127,1135,1143],{"type":45,"tag":246,"props":1120,"children":1121},{"class":248,"line":249},[1122],{"type":45,"tag":246,"props":1123,"children":1124},{},[1125],{"type":51,"value":1126},"for await (const chunk of llmStream) {\n",{"type":45,"tag":246,"props":1128,"children":1129},{"class":248,"line":258},[1130],{"type":45,"tag":246,"props":1131,"children":1132},{},[1133],{"type":51,"value":1134},"    ws.send(JSON.stringify({ type: \"text\", token: chunk, last: false }));\n",{"type":45,"tag":246,"props":1136,"children":1137},{"class":248,"line":267},[1138],{"type":45,"tag":246,"props":1139,"children":1140},{},[1141],{"type":51,"value":1142},"}\n",{"type":45,"tag":246,"props":1144,"children":1145},{"class":248,"line":277},[1146],{"type":45,"tag":246,"props":1147,"children":1148},{},[1149],{"type":51,"value":1150},"ws.send(JSON.stringify({ type: \"text\", token: \"\", last: true }));\n",{"type":45,"tag":758,"props":1152,"children":1154},{"id":1153},"voice-configuration",[1155],{"type":51,"value":1156},"Voice Configuration",{"type":45,"tag":54,"props":1158,"children":1159},{},[1160],{"type":45,"tag":112,"props":1161,"children":1162},{},[1163],{"type":51,"value":1072},{"type":45,"tag":60,"props":1165,"children":1167},{"className":238,"code":1166,"language":240,"meta":69,"style":69},"connect.conversation_relay(\n    url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n    voice=\"en-US-Neural2-F\",\n    language=\"en-US\",\n    transcription_provider=\"deepgram\",\n    speech_model=\"nova-2-phonecall\",\n    interrupt_by_dtmf=True,\n)\n",[1168],{"type":45,"tag":67,"props":1169,"children":1170},{"__ignoreMap":69},[1171,1179,1187,1195,1203,1211,1219,1227],{"type":45,"tag":246,"props":1172,"children":1173},{"class":248,"line":249},[1174],{"type":45,"tag":246,"props":1175,"children":1176},{},[1177],{"type":51,"value":1178},"connect.conversation_relay(\n",{"type":45,"tag":246,"props":1180,"children":1181},{"class":248,"line":258},[1182],{"type":45,"tag":246,"props":1183,"children":1184},{},[1185],{"type":51,"value":1186},"    url=\"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n",{"type":45,"tag":246,"props":1188,"children":1189},{"class":248,"line":267},[1190],{"type":45,"tag":246,"props":1191,"children":1192},{},[1193],{"type":51,"value":1194},"    voice=\"en-US-Neural2-F\",\n",{"type":45,"tag":246,"props":1196,"children":1197},{"class":248,"line":277},[1198],{"type":45,"tag":246,"props":1199,"children":1200},{},[1201],{"type":51,"value":1202},"    language=\"en-US\",\n",{"type":45,"tag":246,"props":1204,"children":1205},{"class":248,"line":286},[1206],{"type":45,"tag":246,"props":1207,"children":1208},{},[1209],{"type":51,"value":1210},"    transcription_provider=\"deepgram\",\n",{"type":45,"tag":246,"props":1212,"children":1213},{"class":248,"line":294},[1214],{"type":45,"tag":246,"props":1215,"children":1216},{},[1217],{"type":51,"value":1218},"    speech_model=\"nova-2-phonecall\",\n",{"type":45,"tag":246,"props":1220,"children":1221},{"class":248,"line":303},[1222],{"type":45,"tag":246,"props":1223,"children":1224},{},[1225],{"type":51,"value":1226},"    interrupt_by_dtmf=True,\n",{"type":45,"tag":246,"props":1228,"children":1229},{"class":248,"line":312},[1230],{"type":45,"tag":246,"props":1231,"children":1232},{},[1233],{"type":51,"value":1234},")\n",{"type":45,"tag":54,"props":1236,"children":1237},{},[1238],{"type":45,"tag":112,"props":1239,"children":1240},{},[1241],{"type":51,"value":1111},{"type":45,"tag":60,"props":1243,"children":1245},{"className":392,"code":1244,"language":394,"meta":69,"style":69},"connect.conversationRelay({\n    url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n    voice: \"en-US-Neural2-F\",\n    language: \"en-US\",\n    transcriptionProvider: \"deepgram\",\n    speechModel: \"nova-2-phonecall\",\n    interruptByDtmf: true,\n});\n",[1246],{"type":45,"tag":67,"props":1247,"children":1248},{"__ignoreMap":69},[1249,1257,1265,1273,1281,1289,1297,1305],{"type":45,"tag":246,"props":1250,"children":1251},{"class":248,"line":249},[1252],{"type":45,"tag":246,"props":1253,"children":1254},{},[1255],{"type":51,"value":1256},"connect.conversationRelay({\n",{"type":45,"tag":246,"props":1258,"children":1259},{"class":248,"line":258},[1260],{"type":45,"tag":246,"props":1261,"children":1262},{},[1263],{"type":51,"value":1264},"    url: \"wss:\u002F\u002Fyourapp.com\u002Fws\u002Fvoice\",\n",{"type":45,"tag":246,"props":1266,"children":1267},{"class":248,"line":267},[1268],{"type":45,"tag":246,"props":1269,"children":1270},{},[1271],{"type":51,"value":1272},"    voice: \"en-US-Neural2-F\",\n",{"type":45,"tag":246,"props":1274,"children":1275},{"class":248,"line":277},[1276],{"type":45,"tag":246,"props":1277,"children":1278},{},[1279],{"type":51,"value":1280},"    language: \"en-US\",\n",{"type":45,"tag":246,"props":1282,"children":1283},{"class":248,"line":286},[1284],{"type":45,"tag":246,"props":1285,"children":1286},{},[1287],{"type":51,"value":1288},"    transcriptionProvider: \"deepgram\",\n",{"type":45,"tag":246,"props":1290,"children":1291},{"class":248,"line":294},[1292],{"type":45,"tag":246,"props":1293,"children":1294},{},[1295],{"type":51,"value":1296},"    speechModel: \"nova-2-phonecall\",\n",{"type":45,"tag":246,"props":1298,"children":1299},{"class":248,"line":303},[1300],{"type":45,"tag":246,"props":1301,"children":1302},{},[1303],{"type":51,"value":1304},"    interruptByDtmf: true,\n",{"type":45,"tag":246,"props":1306,"children":1307},{"class":248,"line":312},[1308],{"type":45,"tag":246,"props":1309,"children":1310},{},[1311],{"type":51,"value":485},{"type":45,"tag":73,"props":1313,"children":1314},{},[],{"type":45,"tag":46,"props":1316,"children":1318},{"id":1317},"cannot",[1319],{"type":51,"value":1320},"CANNOT",{"type":45,"tag":83,"props":1322,"children":1323},{},[1324,1342,1366,1376,1394,1412,1422,1432,1442,1475,1493,1503,1513,1529],{"type":45,"tag":87,"props":1325,"children":1326},{},[1327,1332,1334,1340],{"type":45,"tag":112,"props":1328,"children":1329},{},[1330],{"type":51,"value":1331},"No raw audio access",{"type":51,"value":1333}," — Text in, text out only. For raw audio, use ",{"type":45,"tag":67,"props":1335,"children":1337},{"className":1336},[],[1338],{"type":51,"value":1339},"\u003CConnect>\u003CStream>",{"type":51,"value":1341}," (Media Streams).",{"type":45,"tag":87,"props":1343,"children":1344},{},[1345,1350,1352,1357,1358,1364],{"type":45,"tag":112,"props":1346,"children":1347},{},[1348],{"type":51,"value":1349},"Cannot mix with Media Streams",{"type":51,"value":1351}," — ",{"type":45,"tag":67,"props":1353,"children":1355},{"className":1354},[],[1356],{"type":51,"value":1339},{"type":51,"value":134},{"type":45,"tag":67,"props":1359,"children":1361},{"className":1360},[],[1362],{"type":51,"value":1363},"\u003CConnect>\u003CConversationRelay>",{"type":51,"value":1365}," are mutually exclusive on the same call. No error — one is silently ignored.",{"type":45,"tag":87,"props":1367,"children":1368},{},[1369,1374],{"type":45,"tag":112,"props":1370,"children":1371},{},[1372],{"type":51,"value":1373},"No custom STT\u002FTTS engines",{"type":51,"value":1375}," — Limited to Deepgram + Google (STT) and Deepgram + Amazon Polly + Google + ElevenLabs (TTS).",{"type":45,"tag":87,"props":1377,"children":1378},{},[1379,1384,1386,1392],{"type":45,"tag":112,"props":1380,"children":1381},{},[1382],{"type":51,"value":1383},"No mid-session voice\u002Fprovider changes",{"type":51,"value":1385}," — Voice and provider are set at TwiML time. Only ",{"type":45,"tag":67,"props":1387,"children":1389},{"className":1388},[],[1390],{"type":51,"value":1391},"language",{"type":51,"value":1393}," can be switched mid-session via WebSocket message.",{"type":45,"tag":87,"props":1395,"children":1396},{},[1397,1402,1404,1410],{"type":45,"tag":112,"props":1398,"children":1399},{},[1400],{"type":51,"value":1401},"No WebSocket auto-reconnection",{"type":51,"value":1403}," — If the WebSocket drops, the call disconnects. Implement recovery via ",{"type":45,"tag":67,"props":1405,"children":1407},{"className":1406},[],[1408],{"type":51,"value":1409},"\u003CConnect action>",{"type":51,"value":1411}," URL.",{"type":45,"tag":87,"props":1413,"children":1414},{},[1415,1420],{"type":45,"tag":112,"props":1416,"children":1417},{},[1418],{"type":51,"value":1419},"Voice only",{"type":51,"value":1421}," — No SMS\u002Fmessaging support. For omnichannel, use Conversation Orchestrator.",{"type":45,"tag":87,"props":1423,"children":1424},{},[1425,1430],{"type":45,"tag":112,"props":1426,"children":1427},{},[1428],{"type":51,"value":1429},"No built-in memory or context",{"type":51,"value":1431}," — BYO conversation history and context management.",{"type":45,"tag":87,"props":1433,"children":1434},{},[1435,1440],{"type":45,"tag":112,"props":1436,"children":1437},{},[1438],{"type":51,"value":1439},"No LLM integration",{"type":51,"value":1441}," — Pure transport layer. You bring your own LLM via the WebSocket server.",{"type":45,"tag":87,"props":1443,"children":1444},{},[1445,1450,1451,1457,1459,1465,1467,1473],{"type":45,"tag":112,"props":1446,"children":1447},{},[1448],{"type":51,"value":1449},"No server-side recording via REST API",{"type":51,"value":1351},{"type":45,"tag":67,"props":1452,"children":1454},{"className":1453},[],[1455],{"type":51,"value":1456},"record:true",{"type":51,"value":1458}," on the Calls API is silently ignored. Must use ",{"type":45,"tag":67,"props":1460,"children":1462},{"className":1461},[],[1463],{"type":51,"value":1464},"\u003CStart>\u003CRecording>",{"type":51,"value":1466}," before ",{"type":45,"tag":67,"props":1468,"children":1470},{"className":1469},[],[1471],{"type":51,"value":1472},"\u003CConnect>",{"type":51,"value":1474}," in TwiML.",{"type":45,"tag":87,"props":1476,"children":1477},{},[1478,1483,1485,1491],{"type":45,"tag":112,"props":1479,"children":1480},{},[1481],{"type":51,"value":1482},"Not PCI compliant with Voice Intelligence v2",{"type":51,"value":1484}," — Do not enable ",{"type":45,"tag":67,"props":1486,"children":1488},{"className":1487},[],[1489],{"type":51,"value":1490},"intelligenceService",{"type":51,"value":1492}," in PCI workflows.",{"type":45,"tag":87,"props":1494,"children":1495},{},[1496,1501],{"type":45,"tag":112,"props":1497,"children":1498},{},[1499],{"type":51,"value":1500},"ElevenLabs requires account enablement",{"type":51,"value":1502}," — Accounts without ElevenLabs access get error 64101. Voice IDs (not human-readable names) are required.",{"type":45,"tag":87,"props":1504,"children":1505},{},[1506,1511],{"type":45,"tag":112,"props":1507,"children":1508},{},[1509],{"type":51,"value":1510},"Cannot use ConversationRelay without onboarding",{"type":51,"value":1512}," — Not available immediately on a new account",{"type":45,"tag":87,"props":1514,"children":1515},{},[1516,1521,1523,1528],{"type":45,"tag":112,"props":1517,"children":1518},{},[1519],{"type":51,"value":1520},"Cannot use non-TLS WebSocket",{"type":51,"value":1522}," — Server must be reachable via ",{"type":45,"tag":67,"props":1524,"children":1526},{"className":1525},[],[1527],{"type":51,"value":159},{"type":51,"value":161},{"type":45,"tag":87,"props":1530,"children":1531},{},[1532,1543,1545,1550],{"type":45,"tag":112,"props":1533,"children":1534},{},[1535,1537],{"type":51,"value":1536},"Cannot stream audio without ",{"type":45,"tag":67,"props":1538,"children":1540},{"className":1539},[],[1541],{"type":51,"value":1542},"last: true",{"type":51,"value":1544}," — Twilio won't play audio until it sees a ",{"type":45,"tag":67,"props":1546,"children":1548},{"className":1547},[],[1549],{"type":51,"value":1542},{"type":51,"value":1551}," token in the stream",{"type":45,"tag":73,"props":1553,"children":1554},{},[],{"type":45,"tag":46,"props":1556,"children":1558},{"id":1557},"next-steps",[1559],{"type":51,"value":1560},"Next Steps",{"type":45,"tag":83,"props":1562,"children":1563},{},[1564,1579],{"type":45,"tag":87,"props":1565,"children":1566},{},[1567,1572,1574],{"type":45,"tag":112,"props":1568,"children":1569},{},[1570],{"type":51,"value":1571},"Place outbound calls:",{"type":51,"value":1573}," ",{"type":45,"tag":67,"props":1575,"children":1577},{"className":1576},[],[1578],{"type":51,"value":177},{"type":45,"tag":87,"props":1580,"children":1581},{},[1582,1587,1588],{"type":45,"tag":112,"props":1583,"children":1584},{},[1585],{"type":51,"value":1586},"Standard IVR without AI:",{"type":51,"value":1573},{"type":45,"tag":67,"props":1589,"children":1591},{"className":1590},[],[1592],{"type":51,"value":1593},"twilio-voice-twiml",{"type":45,"tag":1595,"props":1596,"children":1597},"style",{},[1598],{"type":51,"value":1599},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1601,"total":1718},[1602,1620,1634,1646,1664,1686,1706],{"slug":1603,"name":1603,"fn":1604,"description":1605,"org":1606,"tags":1607,"stars":28,"repoUrl":29,"updatedAt":30},"accessibility-and-inclusive-visualization","make data visualizations accessible","Make data visualizations accessible and inclusive. Use when the user needs chart or diagram accessibility guidance, text alternatives for complex visuals, color and contrast review, keyboard support, reduced-motion behavior for animation or parallax, or an accessibility QA workflow for exported figures, UML-like diagrams, and dashboards.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1608,1611,1614,1617],{"name":1609,"slug":1610,"type":15},"Accessibility","accessibility",{"name":1612,"slug":1613,"type":15},"Charts","charts",{"name":1615,"slug":1616,"type":15},"Data Visualization","data-visualization",{"name":1618,"slug":1619,"type":15},"Design","design",{"slug":1621,"name":1621,"fn":1622,"description":1623,"org":1624,"tags":1625,"stars":28,"repoUrl":29,"updatedAt":1633},"agent-browser","automate browser interactions for agents","Browser automation CLI for AI agents. Use when the user needs to interact with websites, verify dev server output, test web apps, navigate pages, fill forms, click buttons, take screenshots, extract data, or automate any browser task. Also triggers when a dev server starts so you can verify it visually.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1626,1627,1630],{"name":17,"slug":18,"type":15},{"name":1628,"slug":1629,"type":15},"Browser Automation","browser-automation",{"name":1631,"slug":1632,"type":15},"Testing","testing","2026-04-06T18:41:03.44016",{"slug":1635,"name":1635,"fn":1636,"description":1637,"org":1638,"tags":1639,"stars":28,"repoUrl":29,"updatedAt":1645},"agent-browser-verify","verify dev server output with automated browser","Automated browser verification for dev servers. Triggers when a dev server starts to run a visual gut-check with agent-browser — verifies the page loads, checks for console errors, validates key UI elements, and reports pass\u002Ffail before continuing.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1640,1641,1644],{"name":1628,"slug":1629,"type":15},{"name":1642,"slug":1643,"type":15},"Local Development","local-development",{"name":1631,"slug":1632,"type":15},"2026-04-06T18:41:17.526867",{"slug":1647,"name":1647,"fn":1648,"description":1649,"org":1650,"tags":1651,"stars":28,"repoUrl":29,"updatedAt":1663},"agents-sdk","build AI agents on Cloudflare Workers","Build AI agents on Cloudflare Workers using the Agents SDK. Load when creating stateful agents, durable workflows, real-time WebSocket apps, scheduled tasks, MCP servers, or chat applications. Covers Agent class, state management, callable RPC, Workflows integration, and React hooks. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1652,1653,1656,1659,1662],{"name":17,"slug":18,"type":15},{"name":1654,"slug":1655,"type":15},"Cloudflare Workers","cloudflare-workers",{"name":1657,"slug":1658,"type":15},"SDK","sdk",{"name":1660,"slug":1661,"type":15},"Serverless","serverless",{"name":13,"slug":14,"type":15},"2026-04-06T18:39:51.717063",{"slug":1665,"name":1665,"fn":1666,"description":1667,"org":1668,"tags":1669,"stars":28,"repoUrl":29,"updatedAt":1685},"ai-elements","build chat UIs with AI Elements","AI Elements component library guidance — pre-built React components for AI interfaces built on shadcn\u002Fui. Use when building chat UIs, message displays, tool call rendering, streaming responses, reasoning panels, or any AI-native interface with the AI SDK.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1670,1673,1676,1679,1682],{"name":1671,"slug":1672,"type":15},"Frontend","frontend",{"name":1674,"slug":1675,"type":15},"React","react",{"name":1677,"slug":1678,"type":15},"shadcn\u002Fui","shadcn-ui",{"name":1680,"slug":1681,"type":15},"UI Components","ui-components",{"name":1683,"slug":1684,"type":15},"Vercel","vercel","2026-04-06T18:40:59.619419",{"slug":1687,"name":1687,"fn":1688,"description":1689,"org":1690,"tags":1691,"stars":28,"repoUrl":29,"updatedAt":1705},"ai-gateway","configure Vercel AI Gateway","Vercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1692,1695,1698,1701,1704],{"name":1693,"slug":1694,"type":15},"AI Infrastructure","ai-infrastructure",{"name":1696,"slug":1697,"type":15},"Cost Optimization","cost-optimization",{"name":1699,"slug":1700,"type":15},"LLM","llm",{"name":1702,"slug":1703,"type":15},"Performance","performance",{"name":1683,"slug":1684,"type":15},"2026-04-06T18:40:44.377464",{"slug":1707,"name":1707,"fn":1708,"description":1709,"org":1710,"tags":1711,"stars":28,"repoUrl":29,"updatedAt":1717},"ai-generation-persistence","implement persistence patterns for AI generations","AI generation persistence patterns — unique IDs, addressable URLs, database storage, and cost tracking for every LLM generation",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1712,1713,1716],{"name":1696,"slug":1697,"type":15},{"name":1714,"slug":1715,"type":15},"Database","database",{"name":1699,"slug":1700,"type":15},"2026-04-06T18:41:08.513425",600,{"items":1720,"total":1917},[1721,1742,1765,1782,1798,1815,1834,1846,1860,1874,1886,1901],{"slug":1722,"name":1722,"fn":1723,"description":1724,"org":1725,"tags":1726,"stars":1739,"repoUrl":1740,"updatedAt":1741},"prior-auth-packet-builder","build healthcare prior authorization packets","Build a concise prior authorization packet from local case files and payer policy docs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1727,1730,1733,1736],{"name":1728,"slug":1729,"type":15},"Documents","documents",{"name":1731,"slug":1732,"type":15},"Healthcare","healthcare",{"name":1734,"slug":1735,"type":15},"Insurance","insurance",{"name":1737,"slug":1738,"type":15},"Regulatory Compliance","regulatory-compliance",28169,"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-agents-python","2026-04-16T05:11:39.180399",{"slug":1743,"name":1743,"fn":1744,"description":1745,"org":1746,"tags":1747,"stars":1762,"repoUrl":1763,"updatedAt":1764},"aspnet-core","build ASP.NET Core web applications","Build, review, refactor, or architect ASP.NET Core web applications using current official guidance for .NET web development. Use when working on Blazor Web Apps, Razor Pages, MVC, Minimal APIs, controller-based Web APIs, SignalR, gRPC, middleware, dependency injection, configuration, authentication, authorization, testing, performance, deployment, or ASP.NET Core upgrades.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1748,1751,1753,1756,1759],{"name":1749,"slug":1750,"type":15},".NET","dotnet",{"name":1752,"slug":1743,"type":15},"ASP.NET Core",{"name":1754,"slug":1755,"type":15},"Blazor","blazor",{"name":1757,"slug":1758,"type":15},"C#","csharp",{"name":1760,"slug":1761,"type":15},"Web Development","web-development",23787,"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fskills","2026-04-12T05:07:02.819491",{"slug":1766,"name":1766,"fn":1767,"description":1768,"org":1769,"tags":1770,"stars":1762,"repoUrl":1763,"updatedAt":1781},"chatgpt-apps","build ChatGPT Apps SDK applications","Build, scaffold, refactor, and troubleshoot ChatGPT Apps SDK applications that combine an MCP server and widget UI. Use when Codex needs to design tools, register UI resources, wire the MCP Apps bridge or ChatGPT compatibility APIs, apply Apps SDK metadata or CSP or domain settings, or produce a docs-aligned project scaffold. Prefer a docs-first workflow by invoking the openai-docs skill or OpenAI developer docs MCP tools before generating code.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1771,1774,1777,1780],{"name":1772,"slug":1773,"type":15},"Apps SDK","apps-sdk",{"name":1775,"slug":1776,"type":15},"ChatGPT","chatgpt",{"name":1778,"slug":1779,"type":15},"MCP","mcp",{"name":9,"slug":8,"type":15},"2026-04-12T05:07:05.468097",{"slug":1783,"name":1783,"fn":1784,"description":1785,"org":1786,"tags":1787,"stars":1762,"repoUrl":1763,"updatedAt":1797},"cli-creator","build CLIs from API docs","Build a composable CLI for Codex from API docs, an OpenAPI spec, existing curl examples, an SDK, a web app, an admin tool, or a local script. Use when the user wants Codex to create a command-line tool that can run from any repo, expose composable read\u002Fwrite commands, return stable JSON, manage auth, and pair with a companion skill.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1788,1791,1794],{"name":1789,"slug":1790,"type":15},"API Development","api-development",{"name":1792,"slug":1793,"type":15},"CLI","cli",{"name":1795,"slug":1796,"type":15},"Codex","codex","2026-04-12T05:07:04.132762",{"slug":1799,"name":1799,"fn":1800,"description":1801,"org":1802,"tags":1803,"stars":1762,"repoUrl":1763,"updatedAt":1814},"cloudflare-deploy","deploy projects to Cloudflare","Deploy applications and infrastructure to Cloudflare using Workers, Pages, and related platform services. Use when the user asks to deploy, host, publish, or set up a project on Cloudflare.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1804,1807,1810,1811],{"name":1805,"slug":1806,"type":15},"Cloudflare","cloudflare",{"name":1808,"slug":1809,"type":15},"Cloudflare Pages","cloudflare-pages",{"name":1654,"slug":1655,"type":15},{"name":1812,"slug":1813,"type":15},"Deployment","deployment","2026-04-12T05:07:14.275118",{"slug":1816,"name":1816,"fn":1817,"description":1818,"org":1819,"tags":1820,"stars":1762,"repoUrl":1763,"updatedAt":1833},"define-goal","define and set measurable project goals","Help the user define a concrete, measurable goal before starting work, especially when they ask to use the goal tool, create a goal, set an objective, clarify success criteria, or turn a fuzzy intention into a quantitative outcome. Use this skill for goal creation and goal refinement only; it does not manage durable snapshots, decision logs, or long-running execution artifacts.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1821,1824,1827,1830],{"name":1822,"slug":1823,"type":15},"Productivity","productivity",{"name":1825,"slug":1826,"type":15},"Project Management","project-management",{"name":1828,"slug":1829,"type":15},"Strategy","strategy",{"name":1831,"slug":1832,"type":15},"Task Management","task-management","2026-05-23T06:17:16.870838",{"slug":1835,"name":1835,"fn":1836,"description":1837,"org":1838,"tags":1839,"stars":1762,"repoUrl":1763,"updatedAt":1845},"figma","translate Figma designs into code","Use the Figma MCP server to fetch design context, screenshots, variables, and assets from Figma, and to translate Figma nodes into production code. Trigger when a task involves Figma URLs, node IDs, design-to-code implementation, or Figma MCP setup and troubleshooting.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1840,1841,1843,1844],{"name":1618,"slug":1619,"type":15},{"name":1842,"slug":1835,"type":15},"Figma",{"name":1671,"slug":1672,"type":15},{"name":1778,"slug":1779,"type":15},"2026-04-12T05:06:47.939943",{"slug":1847,"name":1847,"fn":1848,"description":1849,"org":1850,"tags":1851,"stars":1762,"repoUrl":1763,"updatedAt":1859},"figma-code-connect-components","connect Figma designs to code components","Connects Figma design components to code components using Code Connect mapping tools. Use when user says \"code connect\", \"connect this component to code\", \"map this component\", \"link component to code\", \"create code connect mapping\", or wants to establish mappings between Figma designs and code implementations. For canvas writes via `use_figma`, use `figma-use`.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1852,1853,1856,1857,1858],{"name":1618,"slug":1619,"type":15},{"name":1854,"slug":1855,"type":15},"Design System","design-system",{"name":1842,"slug":1835,"type":15},{"name":1671,"slug":1672,"type":15},{"name":1680,"slug":1681,"type":15},"2026-05-10T05:59:52.971881",{"slug":1861,"name":1861,"fn":1862,"description":1863,"org":1864,"tags":1865,"stars":1762,"repoUrl":1763,"updatedAt":1873},"figma-create-design-system-rules","generate design system rules from Figma","Generates custom design system rules for the user's codebase. Use when user says \"create design system rules\", \"generate rules for my project\", \"set up design rules\", \"customize design system guidelines\", or wants to establish project-specific conventions for Figma-to-code workflows. Requires Figma MCP server connection.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1866,1867,1868,1871,1872],{"name":1618,"slug":1619,"type":15},{"name":1854,"slug":1855,"type":15},{"name":1869,"slug":1870,"type":15},"Documentation","documentation",{"name":1842,"slug":1835,"type":15},{"name":1671,"slug":1672,"type":15},"2026-05-16T06:07:47.821474",{"slug":1875,"name":1875,"fn":1876,"description":1877,"org":1878,"tags":1879,"stars":1762,"repoUrl":1763,"updatedAt":1885},"figma-implement-design","translate Figma designs into application code","Translates Figma designs into production-ready application code with 1:1 visual fidelity. Use when implementing UI code from Figma files, when user mentions \"implement design\", \"generate code\", \"implement component\", provides Figma URLs, or asks to build components matching Figma specs. For Figma canvas writes via `use_figma`, use `figma-use`.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1880,1881,1882,1883,1884],{"name":1618,"slug":1619,"type":15},{"name":1842,"slug":1835,"type":15},{"name":1671,"slug":1672,"type":15},{"name":1680,"slug":1681,"type":15},{"name":1760,"slug":1761,"type":15},"2026-05-16T06:07:40.583615",{"slug":1887,"name":1887,"fn":1888,"description":1889,"org":1890,"tags":1891,"stars":1762,"repoUrl":1763,"updatedAt":1900},"hatch-pet","create animated pets for Codex","Create, repair, validate, visually QA, and package Codex-compatible animated pets and pet spritesheets from character art, generated images, company or prospect brand cues, or visual references. Use when a user wants a lightweight-worker Codex pet workflow, a non-pixel custom pet style, a prospect or company mascot pet, or a full 8x9 animated pet atlas with transparent unused cells, QA contact sheets, and pet.json packaging. This skill composes the installed $imagegen system skill for visual generation and uses bundled scripts for deterministic spritesheet assembly.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1892,1895,1896,1899],{"name":1893,"slug":1894,"type":15},"Animation","animation",{"name":1795,"slug":1796,"type":15},{"name":1897,"slug":1898,"type":15},"Creative","creative",{"name":1618,"slug":1619,"type":15},"2026-05-02T05:31:48.48485",{"slug":1902,"name":1902,"fn":1903,"description":1904,"org":1905,"tags":1906,"stars":1762,"repoUrl":1763,"updatedAt":1916},"imagegen","generate and edit raster images","Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG\u002Fvector\u002Fcode-native assets, extending an established icon or logo system, or building the visual directly in HTML\u002FCSS\u002Fcanvas.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1907,1908,1909,1912,1915],{"name":1897,"slug":1898,"type":15},{"name":1618,"slug":1619,"type":15},{"name":1910,"slug":1911,"type":15},"Image Generation","image-generation",{"name":1913,"slug":1914,"type":15},"Images","images",{"name":9,"slug":8,"type":15},"2026-05-15T06:23:24.312127",675]