This is a simple Telegram bot that receives audio files and converts them to text using the Google speech recognition.
The libraries used are python-telegram-bot for the bot creation, SpeechRecognition for the speech to text conversion and ffmpy to convert the audio to wav.
To make this bot work, you need to set a TOKEN variable in the bot_config.py file. Get this token creating a telegram bot with @BotFather.
In the bot_config.py file, you can set the language that the bot will use to convert the audio to text. This will also be the language that the bot will use to reply to the user. By default there's es-ES for Spanish and en-US for English. If you use a different one, Google will probably recognise it, but the bot will still reply in english unless you create a new texts file for the language.
Since this is using the Google speech recognition service, their limitations apply. There's a 20 MB limit for the audio files (after conversion to wav) so if they're longer than 2 minutes it probably won't work.