AI · 14 — Voice chat on a web page

{
  "schemaVersion": "flowdrome.workflow.v1",
  "id": "",
  "name": "AI · 14 — Voice chat on a web page",
  "version": "0",
  "status": "draft",
  "nodes": [
    {
      "id": "chat",
      "type": "input.webchat",
      "label": "Web Chat",
      "category": "trigger",
      "config": {
        "path": "/chat/ai-demo",
        "title": "Voice assistant",
        "greeting": "Hi! Type, or press the mic and talk — I answer out loud.",
        "placeholder": "Say something…",
        "accent": "#635bff",
        "position": "right",
        "replyField": "",
        "allowedOrigins": "*"
      },
      "position": {
        "x": 0,
        "y": 0
      }
    },
    {
      "id": "hear",
      "type": "ai.transcribe",
      "label": "Transcribe (voice turns)",
      "category": "ai",
      "config": {
        "baseUrl": "http://localhost:8970/v1",
        "apiKey": "",
        "model": "Systran/faster-whisper-tiny.en",
        "audioField": "audio",
        "language": "",
        "outputField": "message",
        "timeoutMs": 120000
      },
      "position": {
        "x": 530,
        "y": 0
      },
      "attachments": {
        "model": "stt"
      }
    },
    {
      "id": "stt",
      "type": "ai.model",
      "label": "Whisper STT (local)",
      "category": "ai",
      "config": {
        "provider": "generic-url",
        "baseUrl": "http://localhost:8970/v1",
        "apiKey": "",
        "model": "Systran/faster-whisper-tiny.en",
        "models": [],
        "temperature": 0.2,
        "maxTokens": 512
      },
      "position": {
        "x": 530,
        "y": 222
      }
    },
    {
      "id": "brain",
      "type": "ai.prompt",
      "label": "Think",
      "category": "ai",
      "config": {
        "provider": "ollama",
        "baseUrl": "",
        "apiKey": "",
        "model": "qwen2.5:1.5b",
        "systemPrompt": "You are a friendly voice assistant. Answer in ONE short sentence — it will be read aloud.",
        "promptMode": "field",
        "promptField": "message",
        "temperature": "0.4",
        "maxTokens": 120,
        "jsonMode": false,
        "timeoutMs": 120000
      },
      "position": {
        "x": 1060,
        "y": 0
      },
      "attachments": {
        "model": "llm"
      }
    },
    {
      "id": "llm",
      "type": "ai.model",
      "label": "Ollama (local)",
      "category": "ai",
      "config": {
        "provider": "ollama",
        "baseUrl": "",
        "apiKey": "",
        "model": "qwen2.5:1.5b",
        "models": [],
        "temperature": 0.2,
        "maxTokens": 512
      },
      "position": {
        "x": 1060,
        "y": 222
      }
    },
    {
      "id": "speak",
      "type": "ai.speak",
      "label": "Speak",
      "category": "ai",
      "config": {
        "baseUrl": "http://localhost:8880/v1",
        "apiKey": "",
        "model": "kokoro",
        "voice": "af_heart",
        "textField": "reply",
        "format": "mp3",
        "timeoutMs": 120000
      },
      "position": {
        "x": 1590,
        "y": 0
      },
      "attachments": {
        "model": "tts"
      }
    },
    {
      "id": "tts",
      "type": "ai.model",
      "label": "Kokoro TTS (local)",
      "category": "ai",
      "config": {
        "provider": "generic-url",
        "baseUrl": "http://localhost:8880/v1",
        "apiKey": "",
        "model": "kokoro",
        "models": [],
        "temperature": 0.2,
        "maxTokens": 512
      },
      "position": {
        "x": 1590,
        "y": 222
      }
    },
    {
      "id": "play",
      "type": "output.audio",
      "label": "Play (editor tests)",
      "category": "output",
      "config": {
        "audioField": "",
        "label": ""
      },
      "position": {
        "x": 2120,
        "y": 0
      }
    }
  ],
  "edges": [
    {
      "id": "e0",
      "sourceNodeId": "chat",
      "sourcePort": "output",
      "targetNodeId": "hear",
      "targetPort": "input"
    },
    {
      "id": "e1",
      "sourceNodeId": "hear",
      "sourcePort": "output",
      "targetNodeId": "brain",
      "targetPort": "input"
    },
    {
      "id": "e2",
      "sourceNodeId": "brain",
      "sourcePort": "output",
      "targetNodeId": "speak",
      "targetPort": "input"
    },
    {
      "id": "e3",
      "sourceNodeId": "speak",
      "sourcePort": "output",
      "targetNodeId": "play",
      "targetPort": "input"
    }
  ],
  "triggers": [
    {
      "id": "t0",
      "nodeId": "chat"
    }
  ],
  "meta": {
    "snippets": [],
    "source": "template-gallery",
    "description": "Everything at once: an embeddable web chat where you can TYPE or TALK (mic button). Voice is transcribed by local Whisper, answered by local Ollama, and spoken back by local Kokoro — a full voice assistant with zero cloud. Deploy it, then open the chat page."
  }
}