How to Run SR-REST Server Image

Speech Recognition

Deploying Image

The application works on 11000 port. Please open the instance through this port.
Instance’s RAM/CPU values that will be used are proportional to the number of licenses and installed language models you purchase. This image has Arabic, Turkish and English language models. It should start with a minimum 4 GB RAM as such. Otherwise, the container will not function properly.

Sestek’s license model is the number of voice recognition actions you perform at the same time.
For example; when you get 5 port licenses, the engine will respond to 5 speech recognition requests at the same time. The license remains closed until the duration of the voice recording you sent.

For 5 ports;
4 GB RAM
2 CPU

Licensing:

Connect the container and run Powershell

Get User ID:

Run the following command to get userid, and send it to Sestek contacts to get License key.

sestek-license.exe userid sr-rest

Register License:

Run the following command for register license.

sestek-license.exe register sr-rest [LICENCE_KEY]

Restart-Service sesteklicensesvc

Now you are ready to make Speech recognition in three languages (ARABIC, TURKISH, ENGLISH)

Running Smoke Test to test all functions:

Run the following executable and see the result.

C:\Sestek\SR-REST\Test\SmokeTests\SpeechRecognition.SmokeTests.exe

First Run

Navigate to following URI to reach API reference doc.

/doc/api

(e.g. http:10.0.0.2:11000/doc/api)

Contact with Sestek

https://support.sestek.com/en-US/

cagri.tezgel@sestek.com

Related Solutions

Text-to-Speech (TTS)
Sestek Text-to-Speech vocalizes texts in real time in a clear and human-like way.

Natural Language Processing
Sestek Natural Language Processing enables human-machine interaction in natural dialog, allowing users to speak to systems in their own words rather than through a menu of fixed phrases.

Related Technologies

Core Technologies
Sestek Core Technologies power different speech solutions that interact with users.