How to hit Language Detection Cognitive Service API through Postman

Question

How to hit Language Detection Cognitive Service API through Postman

sam 201

Hi Team,

I am running Language Detection Cognitive Service API locally.

![108493-aaa.png ][2]

Issue: Getting 404 when i am hitting through Postman

Note: My cognitive service region is westeurope and Pricing tier F0

Answer 1

romungi-MSFT 30,291 Microsoft Employee

@sam I think for a locally run language detection container the detection only works with websockets i.e the SDK.

This is mentioned in the FAQ section of the speech containers. Although language detection status and ready API's seem to support REST API calls.

You can test your local container by running antsu language detection client and check if your local endpoint is working. This command should work in this case.

docker run --rm -ti antsu/on-prem-client:latest ./speech-to-text-with-languagedetection-client ./audio/LanguageDetection_en-us.wav --host localhost --lport 5003 --windows

The error with speech to text recognition is because i am not running speech to text container and not passing sport=5000 in the above command.

sam 201 Reputation points

2021-06-28T07:55:42.753+00:00

Thanks for your reply . can you please help me how to hit this language detection through Postman?
romungi-MSFT 30,291 Reputation points Microsoft Employee

2021-06-28T08:07:24.533+00:00

@sam As mentioned in the swagger document or demo page the detection endpoint is enabled for web socket only. So, I do not think this will work with direct REST API call via postman.
sam 201 Reputation points

2021-06-29T10:52:40.1+00:00

language detection client antsu is working fine locally but my worry is how to setup this on kubernetes will this client takes dynamic audio files from UI and do we have any other client to check language detection?
TIA
romungi-MSFT 30,291 Reputation points Microsoft Employee

2021-06-29T13:23:10.413+00:00
antsu client is 3rd party and can be used for verification purpose to test speech containers as mentioned in its documentation. If you are looking to use the speech container and query the endpoint then the recommendation is to use the SDK and point the endpoint to the local endpoint instead of cloud speech service. In this case:

speech_config = speechsdk.SpeechConfig( endpoint="ws://localhost:5003/speech/languagedetection/cognitiveservices/v1"

For language detection as per the documentation turn off the following property.

speech_config.set_service_property( name='speechcontext-phraseDetection.Mode', value='None', channel=speechsdk.ServicePropertyChannel.UriQueryParameter )

The rest should be similar to using language identification feature from the SDK.

How to hit Language Detection Cognitive Service API through Postman

0 additional answers

Activity