Use this page to test the ELSA unscripted API:
- first record an audio file locally on your computer. In order to get good analysis results try to speak for more than 20s. IF you do not know what to say, you can download this sample audio.
- upload it using the upload button below, together with the API key
Submit your data by clicking on the submit button. Results will be shown in the text box below, in json format.
You can test the same command in a command line with cURL by copy-pasting the command automatically generated above the submit button.
Important: The audio needs to be encoded in the right format in order to be accepted by the service. We currently accept FLAC, OGG-Opus and WAV formats. For WAV, we currently accept the audio in 16bit, 16KHz, mono channel or else the server will not accept them. If you have audios in another format there is a good chance that you will be able to convert them using the ffmpeg tool or, more simply, the sollowing sox command:
sox input_audio_file -r 16000 -c 1 -b 16 audio.wav
NOTE: This version of the API might be slow to respond when the audio is very long. If your use-case involved a long audio recording and streaming is an option, we recommend that you integrate with our API in streaming mode (ask us how and we will assist).