Save a mono 16-bit WAV copy of a recording at a lower sample rate. For speech analysis, 16 kHz is plenty: a 16 kHz mono file is far smaller and uses far less memory when analysed in the browser, especially on a phone. Use the saved copy as the input for everything, the web pages and the command-line tools alike, so they all analyse the same samples. Everything happens in your browser; nothing is uploaded.
Settings
A mono file is saved as it is, whatever the channel setting. Channel 1 is what the analysis tools read from a multichannel file. Audio already at or below the chosen rate is not upsampled; it is saved at its own rate.
Files
Files are converted one at a time. WAV files are read at their own rate; other formats are decoded by the browser straight to the chosen rate.