How to use tesseract OCR to recognize image text not from a file, but from a buffer in a node?

Question

JavaScript

There is a module https://www.npmjs.com/package/node-tesseract, everything works fine with it, but I did not find the ability to recognize from the buffer. Then, I found the https://www.npmjs.com/package/tesseract-wrapper module , and I had difficulties with it, as I understood I needed an image in base64 format, I managed to do it like this:

the error received: