Speech Recognition

Whisper

Automatic speech recognition system with multilingual support

Whisper is OpenAI's robust automatic speech recognition system trained on diverse audio data from the web. It supports transcription in multiple languages and can handle various audio qualities and accents with remarkable accuracy. The model is particularly useful for transcription services, accessibility applications, and multilingual communication tools.

Features

Languages

99+ supported languages

Audio Quality

Robust to noise and accents

Model Sizes

Multiple variants available

Open Source

Freely available for use

Layout example

Add the widget script to your layout for multiple pages.

File:components/SensayScript.tsx

"use client";

import Script from "next/script";
import { useEffect, useId } from "react";

export default function SensayScript({ src }: { src: string }) {
  // Generate a unique ID for script so NextJS can run it on client side navigation.
  // NextJS Script component by default runs script only once based on the src value.
  const id = useId();
  const url = new URL(src);
  url.searchParams.set("id", id);

  // Destroy the chatbot when the component unmounts.
  // Destroy chatbot elements as NextJS doesn't clean up the Script component on client side navigation.
  useEffect(() => {
    return () => {
      window.SensayChatbot?.destroy();
    };
  }, []);

  return <Script src={url.toString()} strategy="afterInteractive" />;
}

File:app/products/layout.tsx

import SensayScript from "@/components/SensayScript";

export default function ProductsLayout({ children }) {
  return (
    <div>
      {children}
      <SensayScript src="https://chat-widget.sensay.io/<chatbot-identifier>/embed-script.js" />
    </div>
  )
}