react-ai-voice-agent-sdk

Site-agnostic AI voice agent SDK for React apps (client-side).
This SDK scans actionable DOM elements, listens to voice (Web Speech API), sends transcript + DOM snapshot to your backend /intent endpoint, and executes the returned action safely.

Install

npm i react-ai-voice-agent-sdk

Usage (React)

import React from "react";
import { useVoiceAgent, VoiceAgentOverlay } from "react-ai-voice-agent-sdk";

export default function App() {
  const { agent, state } = useVoiceAgent({
    intentEndpoint: "http://localhost:4000/intent",
    wakeWord: "hey assistant",
    permissions: {
      allowNavigation: false,
      requireConfirmFor: ["delete", "logout", "pay", "checkout"]
    }
  });

  return (
    <>
      <YourApp />
      <VoiceAgentOverlay agent={agent} state={state} />
    </>
  );
}

Backend contract

POST /intent

{
  "text": "hey assistant click login",
  "domSnapshot": [{ "id":"el-1", "role":"button", "label":"Login", "selector":"#login" }]
}

Returns:

{
  "action": "click",
  "targetId": "el-1",
  "value": null,
  "confidence": 0.72
}