How to make Whisper API not fix the errors (give just the RAW voice input)?

Using the following code. Found, that it automatically fix grammar errors, which user done during saying voice message:

require('dotenv').config()
const { Configuration, OpenAIApi } = require("openai");

const path = require("path");
const fs = require("fs")

const convertTransScript = async () => {
    const filePath = path.join(__dirname, 'sample.mp3');

    const configuration = new Configuration({
        apiKey: process.env.OPEN_AI_APK_KEY
    });

    const openai = new OpenAIApi(configuration);
    const resp = await openai.createTranscription(
        fs.createReadStream(filePath),
        "whisper-1"
    ).catch((resError) => {
        console.error(JSON.stringify(resError))
    })

    console.log("RES",resp)
}
convertTransScript();




How to adjust it to make Whisper API not fix the errors (give just the RAW voice input)? Thank you

1 Like