WinForms Chat Speech-to-Text Integration
RadChat provides built-in speech-to-text functionality that allows users to dictate messages using their microphone. The feature leverages the RadSpeechToTextButtonElement component, which is integrated into the chat input area through the RadPromptInputElement.
The speech-to-text feature requires the WebView2 runtime to be installed on the end user machine. Ensure that your application includes the
Microsoft.Web.WebView2NuGet package as a dependency.
Overview
The speech-to-text button is positioned in the input area buttons panel, alongside the send button. When clicked, it activates the speech recognizer which listens for voice input and inserts the recognized text into the chat input text box. The recognized text is inserted at the current caret position, allowing users to combine typed and dictated input seamlessly.
The integration supports:
- Continuous recognition mode — the recognizer listens continuously until explicitly stopped.
- Single utterance mode — the recognizer stops automatically after a pause in speech.
- Configurable language — specify the recognition language via a BCP 47 language tag.
- Visual state feedback — the button animates while actively listening.
Removing the Speech-to-Text Button
The speech-to-text button visibility is controlled by the IsSpeechToTextButtonVisible property on RadChat. By default, this property is set to true.
Toggling the speech-to-text button visibility
this.radChat1.IsSpeechToTextButtonVisible = false;Configuring the Speech Recognizer
You can configure the speech recognition behavior through the RadSpeechToTextButtonElement exposed by the prompt input element. Access it via the RadChat.ChatElement.PromptInputElement.ButtonsPanelElement.SpeechToTextButton property path.
Language Tag
Set the LanguageTag property to specify the recognition language using an IETF BCP 47 tag (for example, "en-US", "de-DE", "fr-FR").
Setting the recognition language
RadSpeechToTextButtonElement sttButton = this.radChat1.ChatElement.PromptInputElement.ButtonsPanelElement.SpeechToTextButton;
sttButton.LanguageTag = "de-DE";Continuous Recognition
The IsContinuousRecognition property determines whether the recognizer listens continuously or stops after a single utterance. The default value is true.
Disabling continuous recognition
RadSpeechToTextButtonElement sttButton = this.radChat1.ChatElement.PromptInputElement.ButtonsPanelElement.SpeechToTextButton;
sttButton.IsContinuousRecognition = false;When continuous recognition is disabled, the recognizer stops automatically after a single utterance or pause in speech.
Events
The RadSpeechToTextButtonElement exposes the following events relevant to the chat integration:
| Event | Description |
|---|---|
| SpeechRecognized | Raised when the speech recognizer successfully recognizes speech. The recognized text is automatically inserted into the input text box. |
| StateChanged | Raised when the recognizer state changes (for example, from Ready to Listening, or from Listening to StoppingListening). |
| ErrorOccurred | Raised when an error is encountered during speech recognition initialization or processing. |
Handling the SpeechRecognized Event
You can subscribe to the SpeechRecognized event to perform custom logic when the recognizer produces a result.
Subscribing to the SpeechRecognized event
RadSpeechToTextButtonElement sttButton = this.radChat1.ChatElement.PromptInputElement.ButtonsPanelElement.SpeechToTextButton;
sttButton.SpeechRecognized += this.OnSpeechRecognized;
private void OnSpeechRecognized(object sender, SpeechRecognizerSpeechRecognizedEventArgs e)
{
// e.Text contains the recognized speech
Console.WriteLine("Recognized: " + e.Text);
}Handling the ErrorOccurred Event
Subscribe to the ErrorOccurred event to handle initialization failures or runtime errors gracefully.
Subscribing to the ErrorOccurred event
RadSpeechToTextButtonElement sttButton = this.radChat1.ChatElement.PromptInputElement.ButtonsPanelElement.SpeechToTextButton;
sttButton.ErrorOccurred += this.OnSpeechToTextError;
private void OnSpeechToTextError(object sender, SpeechRecognizerErrorOccurredEventArgs e)
{
RadMessageBox.Show("Speech recognition error: " + e.ErrorMessage);
}Speech Recognizer States
The button reflects the current state of the speech recognizer through visual feedback. The following table lists all possible states:
| State | Description |
|---|---|
| NotInitialized | The recognizer has not been initialized yet. |
| Ready | The recognizer is initialized and ready to start listening. |
| StartingListening | The recognizer is transitioning to the listening state. |
| Listening | The recognizer is actively listening for speech input. The button shows a pulse animation. |
| StoppingListening | The recognizer is transitioning from listening to ready. |
| Faulted | An error occurred during recognition. |
Requirements
- The
Microsoft.Web.WebView2NuGet package must be referenced by the project. - The WebView2 runtime must be available on the target machine.
- A working microphone must be connected and accessible to the application.