Starting the mercuryScribe Web Server

Using Docker (Mac, Linux, or Windows)

To start the mercuryScribe web server (you need Docker installed), open a terminal and run the following command:

docker run --rm -p 5001:5001 -v "$(pwd)/mercuryscribe:/app/mercuryscribe" ghcr.io/literatecomputing/transcribe-with-whisper-web:latest

Using the Windows Package

If you installed mercuryScribe using the Windows Package, just unzip the package and double-click the start-server.bat file to start the server. Closing this window will stop the server! If you do close it, just double-click the start-server.bat file again to restart the server.

Accessing the Web Interface

Wait a few minutes for the server to download and start. Then open http://localhost:5001 in your web browser.

You will be prompted to provide a Hugging Face Access Token with instructions for obtaining one. There are links for creating a token and accepting model terms, but if that sounds hard, watch the video below. There are also instructions on the page where you enter the token, you might need to scroll up to see them.

You need to accept the model terms before using the API. Most models ask for a brief description of your intended use case so that they can track usage. This is a requirement for using the API.

Get Started with the CLI

If you have many files to transcribe, you might find the command line version. Using Docker will get you all of the required dependencies installed quickly, but if you're comfortable with Python, you can install via pip.

Docker

docker run -v $(pwd):/data literatecomputing/transcribe-with-whisper input.mp4

Python

pip install transcribe-with-whisper

ffmpeg is required; see the Github repo for details.