You can improve the accuracy of Rev.ai transcripts by submitting a list of custom vocabulary terms. This feature is available for both our offline (async) transcription API and our live streaming API.
Who should use this feature?
You should use this feature if your audio contains truly unique or rare terms that you believe are not in our massive dictionary of 500k+ words.
- Examples of good custom vocabulary terms include made-up words or words with unique spelling such as: sparkletini, timi, Ginnifer
- Examples of bad custom vocabulary terms that you should not include are: Maybelline, e pluribus unum, orthostatic hypotension
How many custom vocabulary terms can I include?
You can create an unlimited number of custom vocabulary groups. Each group can have its own list of phrases.
We recommend submitting a short list of target terms (no more than 500 phrases) as large lists may negatively impact performance and accuracy. For example, submitting a massive list can cause our speech engine to detect those words too often and/or in the wrong places. Short phrases also do better than long phrases, so keep your phrases on the short side if possible.
Great, now how do I implement this?
For detailed steps on technical implementation: