pix2txt

Optical Character Recognition (OCR) telegram bot built using python that extracts texts found inside pictures.

usage

1. /start the bot @pix2txt_bot
2. Just send a picture to the bot as you normally would without a caption -- captions & any other texts outside of the picture are unnecessary.
3. It also supports forwarded picture messages.
4. If you want further help, just send /help

libraries employed

• ocrspace
• telethon

how to run

1. create a virtual environment, source it and install -- upgrade if you wish -- all dependencies found inside requirements.txt by doing pip install -U -r requirements.txt

2. get your own API_ID & API_HASH from https://my.telegram.org/auth

3. create a telegram bot using @botfather

4. get bot's token from @botfather

5. create a .env file in the root directory and place the above details inside it as such:

    TOKEN = xxxxxxx
    API_ID = xxxxxx
    API_HASH = xxxxxxxx
   

6. run it just as any python application by python src/botter.py

Caveats

It only recognizes LATIN characters.