Inspiration
Not wanting to listen to audio messages.
What it does
Detects audio messages and transcribes them into text. Also has options for filtering profane language and summarising long messages (more than 100 words long).
How we built it
Using PyTelegramBotAPI and the Huggingface API.
Challenges we ran into
The hugging face model (openai/whisper-large-v3) couldn't handle audio messages longer than ~30 seconds, so we needed to slice the message and pass each slices one by one.
Accomplishments that we're proud of
Getting the bot to work and be somewhat clean.
What we learned
How to use the huggingface api.
What's next for Cookies&audios
sleep
Built With
- huggingface
- python
- telebot
Log in or sign up for Devpost to join the conversation.