It often happens that you cannot listen to a voice message, but the information in it is very important? This bot will help you. The bot turns a voice message in to a text message.
After sending a voice message, the bot downloads it in ogg format. The speech recognition library does not work with this format, so first the bot will convert the received file in to a wav format, using FFmpeg. After that, the speech recognition function is launched. Heard text is written to a variable and sent to the user. If in the message a silence or the words are illegible - the user will get an error. After the end of the work, the bot automatically delete all unnecessary files.
There are 2 possibilities to install and use - Docker and Local
First of all, install Docker and Docker Compose. Download the git repository on your PC. Create your own Telegram Bot using BotFather and get a token, which should be used as 'BOT_TOKEN' variable. Change file name example.env to .env and insert your token there and then run the following command:
docker compose upFirst of all, you need to download the repository. And use this command to install the requirements:
pip install -r requirements.txtFor the correct work you also need a FFmpeg. You can install it as a plugin, use a library, or use a buildpack.
| FFmpeg | Link |
|---|---|
| Plugin | https://www.ffmpeg.org/ |
| Library | https://github.com/kkroening/ffmpeg-python |
| Buildpack | https://github.com/jonathanong/heroku-buildpack-ffmpeg-latest |