{
  "id": "stack-profile:voice-ai-agent",
  "_kind": "StackProfile",
  "_file": "domain/stack-profiles/deep-stacks-1.yaml",
  "_cluster": "domain",
  "attributes": {
    "displayName": "Voice AI Agent Stack (Whisper, TTS, WebSocket, FastAPI, React)",
    "description": "An end-to-end voice-powered AI agent architecture for building\nconversational interfaces with speech input and output. OpenAI Whisper\n(or whisper.cpp) handles automatic speech recognition, converting audio\nstreams to text. A text-to-speech engine synthesizes agent responses\nback to audio. WebSocket connections enable full-duplex, low-latency\naudio streaming between client and server.\n\nFastAPI serves as the async backend, coordinating ASR, LLM inference,\nand TTS in a streaming pipeline. React powers the frontend with audio\ncapture, playback, and visual feedback. Python handles all server-side\nlogic including audio preprocessing and LLM integration. This stack\nsuits voice assistants, call center copilots, and accessibility-first\napplications. The main tradeoff is latency — the ASR-to-TTS round trip\nmust stay under 1-2 seconds for natural conversation flow.\n",
    "composes": [
      "framework:fastapi",
      "framework:react",
      "language:python",
      "language:typescript",
      "library:websockets"
    ]
  },
  "outgoingEdges": [
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "framework:fastapi",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "framework:react",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "language:python",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "language:typescript",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "library:websockets",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "tool:docker",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "library:uvicorn",
      "kind": "composed_of"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "role:ml-engineer",
      "kind": "used_by_role"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "role:fullstack-engineer",
      "kind": "used_by_role"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "role:frontend-engineer",
      "kind": "used_by_role"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "workflow:prompt-engineering-iteration",
      "kind": "follows_workflow"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "workflow:agent-evaluation-cycle",
      "kind": "follows_workflow"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "domain:ml-ai",
      "kind": "applies_to"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "domain:frontend",
      "kind": "applies_to"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "skill-area:audio-processing",
      "kind": "requires_skill_area"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "skill-area:streaming-realtime-processing",
      "kind": "requires_skill_area"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "skill-area:websocket-design",
      "kind": "requires_skill_area"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "skill-area:natural-language-processing",
      "kind": "requires_skill_area"
    },
    {
      "from": "stack-profile:voice-ai-agent",
      "to": "skill-area:model-serving-deployment",
      "kind": "requires_skill_area"
    }
  ],
  "incomingEdges": []
}