No direct answer on voice message

[Marcello87]

I first thought it was a error.
Would it be possible to answer (with chatgpt) directly after whisper has transcribe the audio message?

This would eliminate the manual step copying of the text after transcibe.

[Edoardo]

Hi @Marcello87, that's interesting. I originally intended it as a way to only transcribe audio messages, and not as a direct voice command to ChatGPT.

Should be as easy as replacing these lines:

chatgpt-telegram-bot/telegram_bot.py

Lines 125 to 131 in 946f6a4

 # Send the transcript 
 await context.bot.send_message( 
     chat_id=chat_id, 
     reply_to_message_id=update.message.message_id, 
     text=transcript, 
     parse_mode=constants.ParseMode.MARKDOWN 
 ) 

With:

response = self.openai.get_chat_response(chat_id=chat_id, query=transcript)
await context.bot.send_message(
    chat_id=chat_id,
    reply_to_message_id=update.message.message_id,
    text=response,
    parse_mode=constants.ParseMode.MARKDOWN
)

But I'm not sure how to allow both modes. Will keep this issue open for now, in case someone wants to help

[Marcello87]

Tested, works great for me, many thanks for the quick response.

[sheepgreen]

I think it's difficult to use both modes at the same time. Maybe we can set a parameter like VOICE_TO_CHAT and let users select their prefer mode.

[Edoardo]

Thanks @Marcello87 @slippersheepig
Added VOICE_REPLY_WITH_TRANSCRIPT_ONLY env parameter which defaults to true

[Dean Xi]

When I set VOICE_REPLY_WITH_TRANSCRIPT_ONLY to false, can I let the bot transcribe my voice and output it first, then reply it? It's more user-friendly.