Telegram bot, that helps transcribe and then sum up content of your voices, audios and meetings.

Install

• clone repo
• install dependances using poetry
• app/bot.py to run service
• there is also dockerfile, but it is mostly for prod

To dos

• limitation:
  • audiofiles more than 25mb
    • chunking
    • silence
    • archiving
  • there is limit (technocal (4096 utf) and cognitive) on size of message sent back -- need to cut it wisely
• beautifying answers
• DevOps:
  • docker
    • refact it to 2 sage build with only venv
    • docker compose
  • CI/CD -- github action on push/merge
  • administration
    • logging
• performanse
  • ffmpeg - takes time for conversion
    • use python instead of cli
  • stream io -- less iops
• social functioanality
  • user accounting
  • statistic
• count channels to expect number of speaking persons