It supports open-ended transcription too: https://voice2json.org/commands.html#open-transcription

Users have reported good accuracy with the English Deepspeech profile: https://github.com/synesthesiam/voice2json-profiles