VoiceMessage Telegram Bot

VoiceMessageBotApp is a Kotlin application that utilizes the OpenAI and Telegram API to interact with users through a Telegram Bot. The primary functionality of the bot includes transcribing voice messages, audio files, and YouTube videos into text. Users can also request a summarized version of the transcribed text.

Features

• Transcribe voice messages and audio files sent to the bot
• Transcribe audio from YouTube video URLs sent to the bot using existing subtitles or by downloading the video and converting it to audio
• Summarize the transcribed text upon user request

• Can be run locally, or as a Docker container
• Uses Dotenv to load environment variables from a .env file, so it can be run locally without having to set environment variables

Prerequisites

• JDK 17 or higher
• Maven
• FFmpeg
• yt-dlp
• Telegram API Token
• OpenAI API Token

Usage

Follow these instructions on how to set up your own a Telegram bot: https://core.telegram.org/bots/features#botfather then, copy the token and chat id from the botfather response and set them as environment variables.

The bot is intended to be run as a docker container. The following environment variables are required for it to work:

• CHAT_ID: The id of the chat to send the messages to
• OPENAPI_TOKEN: The token of the Whisper API
• TELEGRAM_BOT_NAME: The name of the telegram bot
• TELEGRAM_BOT_TOKEN: The token of the telegram bot
• FFMPEG_PATH The path to the ffmpeg executable
• YTDLP_PATH The path to the yt-dlp executable

Building

For local usage

The bot can be built using maven. The following command will build an app.jar file in the target directory:

mvn clean package

Feel free to place the following blank .env file in the root directory of the project and adjust to your needs:

CHAT_ID=MyTelegramChatId
TELEGRAM_BOT_NAME=MyTelegramBotName
TELEGRAM_BOT_TOKEN=MyTelegramBotToken
OPENAPI_TOKEN=MyOpenApiToken
FFMPEG_PATH=/usr/local/bin/ffmpeg
YTDLP_PATH=/usr/local/bin/yt-dlp

Or simply set the environment variables in the OS of your choice and run the jar file:

java -jar target/app.jar

For Docker

The bot can also be built as a docker container. The following command will build a docker image called voicemessage-bot:

docker build -t voicemessage-bot .

Place these environment variables in the Dockerfile and adjust to your needs:

ENV CHAT_ID=MyTelegramChatId
ENV TELEGRAM_BOT_NAME=MyTelegramBotName
ENV TELEGRAM_BOT_TOKEN=MyTelegramBotToken
ENV OPENAPI_TOKEN=MyOpenApiToken
ENV FFMPEG_PATH=/usr/local/bin/ffmpeg
ENV YTDLP_PATH=/usr/local/bin/yt-dlp

Disclaimer

This bot is not affiliated with Telegram, OpenAI or YouTube in any way. It is a research project and not intended to be used in any way that violates either company's ToS.