Convenient and decent quality Used Temi on my recording. Good value for the price. The quality was good enough to skim the doc and get the gist of what was being said. I was impressed it got some names of some people right. Considering Siri can't even understand what I'm saying at times this was better than I expected, lol

Not sure if this is what I should expect going forward... I've been using Temi for years--and up to the last couple of recent transcriptions, I've been satisfied. Not so with the last couple. Because there are always some mispronounced words, it is pretty common for me to work the transcript while occasionally listening to a specific part of the recording for a word or short phrase. Not so with the last couple of transcripts. It missed major parts of sentences that were clearly annunciated and identifiable when I listened to the recording. Accuracy was highly affected. The back and forth between transcript and listening significantly slowed me down. In short, I did not trust the transcription. I'll give it a few more shots--but if what I experienced in this last transcript is an indication of where the program is headed, it won't be too long before I find a better option.

This is my first time trying this tool This is my first time trying this tool, and as a Pastor who speaks every week, I'm often listening to audio resources for ideas and inspiration. Taking those inspirations from Audio to an editable transcript in a matter of moments has saved me several hours this week alone! Simple to use and delivers what it promises! I'm hooked!

I am impressed with how well most words were transcribed I am impressed with how well most words were transcribed, given the background noise. And that people were speaking from many different directions. It breaks down in two categories: 1) I am surprised that it does not notice the switch between obviously different speakers. 2) There were a number of proper names and obscure words where it makes amusing guesses. Example: "Bayesian" became "Beijing". I have no idea how that could be improved without knowing a lot more about the subject under discussion.

Cool idea, needs a little more machine-learning along the way Super-cool idea, and way cheaper than person-based transcription, but def takes effort to review/edit the transcription (though less than fully transcribing the original audio). Would be better if there was account-specific machine learning, so that I could upload a reviewed document and have the system learn from its initial mistakes. (e.g., the acroynym NASDDDS kept coming out as "nasty's" -- that would be an easy fix).