Skip to content

Word Error Rate Tool

Measure automatic speech recognition accuracy objectively

human

Step 1

Create a manual transcript of audio recording with 99%+ accuracy

settings

Step 2

Create an automatically generated transcript through speech recognition

copy

Step 3

Paste manual transcript in box 1, paste automatic transcript in box 2


Step 4

Click ‘Calculate’ and check the Word Error Rate

Product mockup
Get your texts

Upload your audio or video file. Our speech recognition engine creates a text, which you can export and paste in our WER tool

Try it free
accuracy
Best accuracy

We build speech recognition engines with the highest accuracy in the market. We are proud of this, which is why we regularly compare our accuracy with competitors. We do this by comparing consistent test sets of audio files with perfect manually-made texts

Learn more
Answers to our

FAQ

Learn more about which solution is best for you.

Contact us
  • What can the WER tool be used for?
    The WER tool can be used to check the accuracy of the transcript that was automatically generated using Amberscript. You will need a 100% accurate version of the transcript as well as the transcript generated by Amberscript’s software. Simply upload the 100% accurate version in the left text box and the transcript generated using Amberscript in the right text box. The WER will be automatically calculated and displayed just above the two text boxes.
  • What is a WER (Word Error Rate)?
    The Word Error Rate (WER) is a measure of the performance of speech recognition technology. It is the percentage of words that is not correctly transformed from speech-to-text. The accuracy is thus 100% minus the WER.

Find your solution here

Products

Learn more about which solution is best for you. Or start uploading your files immediately.

Start now