この記事を参考にすると、YouTube Live チャットのコメントで特定の文字を取得してラズパイでLチカできる。
https://qiita.com/iroiro_bot/items/ad0f3901a2336fe48e8f
↓こんな感じ。
ylc.py
#　Live動画のコメント取得

import time
import requests
import json
import RPi.GPIO as GPIO
import time

#事前に取得したYouTube API key
YT_API_KEY = ''


def get_chat_id(yt_url):
    '''
    https://developers.google.com/youtube/v3/docs/videos/list?hl=ja
    '''
    video_id = yt_url.replace('https://www.youtube.com/watch?v=', '')
    print('video_id : ', video_id)

    url    = 'https://www.googleapis.com/youtube/v3/videos'
    params = {'key': YT_API_KEY, 'id': video_id, 'part': 'liveStreamingDetails'}
    data   = requests.get(url, params=params).json()

    liveStreamingDetails = data['items'][0]['liveStreamingDetails']
    if 'activeLiveChatId' in liveStreamingDetails.keys():
        chat_id = liveStreamingDetails['activeLiveChatId']
        print('get_chat_id done!')
    else:
        chat_id = None
        print('NOT live')

    return chat_id




def get_chat(chat_id, pageToken, log_file):
    '''
    https://developers.google.com/youtube/v3/live/docs/liveChatMessages/list
    '''
    url    = 'https://www.googleapis.com/youtube/v3/liveChat/messages'
    params = {'key': YT_API_KEY, 'liveChatId': chat_id, 'part': 'id,snippet,authorDetails'}
    if type(pageToken) == str:
        params['pageToken'] = pageToken

    data   = requests.get(url, params=params).json()

    try:
        for item in data['items']:
            channelId = item['snippet']['authorChannelId']
            msg       = item['snippet']['displayMessage']
            usr       = item['authorDetails']['displayName']
            #supChat   = item['snippet']['superChatDetails']
            #supStic   = item['snippet']['superStickerDetails']

            # コメントが、Hogeの場合
            if msg == "Hoge":
                print('\n')
                print("Hogeをありがとうございます")
                print('\n')

                #LEDをチカチカ
                PNO = 4 # 抵抗に繋いだ側の、GPIOポート番号
                GPIO.setmode(GPIO.BCM)
                GPIO.setup(PNO, GPIO.OUT)
                for i in range(5):
                    GPIO.output(PNO, GPIO.HIGH) # 点灯
                    time.sleep(0.3)
                    GPIO.output(PNO, GPIO.LOW) # 消灯
                    time.sleep(0.3)
                GPIO.cleanup()


            log_text  = '[by {}  https://www.youtube.com/channel/{}]\n  {}'.format(usr, channelId, msg)
            with open(log_file, 'a') as f:
                print(log_text, file=f)
                print(log_text)
        print('start : ', data['items'][0]['snippet']['publishedAt'])
        print('end   : ', data['items'][-1]['snippet']['publishedAt'])

    except:
        pass

    return data['nextPageToken']




def main(yt_url):
    slp_time        = 10 #sec　終了予定
    iter_times      = 60 #回
    take_time       = slp_time / 60 * iter_times
    print('{}分後　終了予定'.format(take_time))
    print('work on {}'.format(yt_url))

    log_file = yt_url.replace('https://www.youtube.com/watch?v=', '') + '.txt'
    with open(log_file, 'a') as f:
        print('{} のチャット欄を記録します。'.format(yt_url), file=f)
    chat_id  = get_chat_id(yt_url)

    nextPageToken = None
    for ii in range(iter_times):
        #for jj in [0]:
        try:
            print('\n')
            nextPageToken = get_chat(chat_id, nextPageToken, log_file)
            time.sleep(slp_time)
        except:
            break




if __name__ == '__main__':
    yt_url = input('Input YouTube URL > ')
    main(yt_url)