CoVoX

Cloud enabled library providing a customizable voice-interface for your application or your device

Covox allows the interaction with an application or device through voice.
You provide a list of Commands, i.e. operations that can be invoked via the voice interface, Covox then listens to the audio and when a command matches with the spoken words, it's executed. It also has multi-language support!

With some imagination you could speak to a calculator, a virtual assistant, or a CRM application!

How it works

1) define commands and provide them to a CovoxEngine instance 2) start the audio capture by calling the method covox.StartAsync 3) covox will translate and recognize the input, and then it will emit the event Recognized 4) execute the logic connected to the detected command

Getting started

Covox is offered as .NET library and acts on behalf of the Azure Cognitive Services, therefore to use it you will need:

• an Azure Cognitive Services subscription key (follow this guideline)
• a .NET project or application
• a device connected to internet
• a device with a working microphone

In order to get started, take a look at the samples.

How to use

Consider a simple use case: a voice-controlled light-switching application.

• Define the available commands, with unique IDs and one or many voice triggers (in English):

var turnOnLightCmd = new Command
{
    Id = "TurnOnLight",
    VoiceTriggers = new[] { "turn on the light", "light on", "on" }
};

var turnOffLightCmd = new Command
{
    Id = "TurnOffLight",
    VoiceTriggers = new[] { "turn off the light", "light off", "off" }
};

• Create an instance of CovoxEngine:

var covox = new CovoxEngine(new Configuration
{
    AzureConfiguration = AzureConfiguration.FromSubscription(
        subscriptionKey: YOUR_SUBSCRIPTION_KEY,
        region: YOUR_REGION),

    // Define all the languages that can be regognized
    InputLanguages = new[] { "en-US", "de-DE", "it-IT", "es-ES" },
});

covox.RegisterCommands(turnOnLightCmd, turnOffLightCmd);

• Define a delegate for when a command is recognized:

covox.Recognized += (cmd, ctx) =>
{
    if (cmd == turnOnLightCmd) { /* ... */ }
    else if (cmd == turnOffLightCmd) { /* ... */ }
};

await covox.StartAsync();

Use case scenarios

AI/Machine Learning

### Guess-Who Guess Who is a game for 2 players. Each player has a "playing field" with different people and a fixed person, which must be guessed by the opponent, by exclusion questions. Via Voice commands you should be able to ask a question, such as, "Does the woman have red hair?" Image recognition should then return the answer yes / no. **Procedure** 1. Asking a Question via Voice Command 2. Recognize and process question 3. Looking at e.g. Image and detect the answer 4. Returning Answer (Yes / No) ### Technologies - CoVoX engine - Python / Tensorflow - [Face](https://azure.microsoft.com/en-us/services/cognitive-services/face/)

---

Technologies

The library is developed in .NET 5 and uses the Azure's Cognitive Services.