I want to programm my own "Alexa" (complete newbie).

[bowrilla]

21 hours ago, Franck said:

If you want an easy starting point you can use C# and windows speech recognition engine. You need about 15 lines of code to get the microphone input converted into text or send text to the speech synthesizer so it output in your speakers. Then 99% of the job is to interpret and generate text.

 

Edit :

Just found the correct class to use and these basic 5 lines of code works to return the spoken audio to text

https://docs.microsoft.com/en-us/dotnet/api/system.speech.recognition.speechrecognitionengine?view=netframework-4.8

 

the "e.Result.Text" in the event at the end of the code is the text of what has been recognized

 

There's a plethora of speech recognition APIs out there and implementations for basically all the common and often also uncommon languages you can come up with.

 

Edit: Here's a python library supporting 8 different speech recognition engines/APIs: https://pypi.org/project/SpeechRecognition/ including MS Bing Voice Recognition. For JS there's an experimental feature in development (Web Speech API) and for nodeJS there's a ton of packages. And here's a repo with a Rust implementation of Mozilla's DeepSpeech library: https://github.com/RustAudio/deepspeech-rs