Tutorials: Voice Widget: Collect Input
Required access: Voxtelesys Portal
The Collect Input widget is one that gets used quite frequently!
You will want to place the answer call widget at the front of the flow if you plan on collecting input/directing the caller around.
The Collect Input widget goes under the voice category for the VAST flow builder. This widget plays an audio file or speaks TTS and then collects a user’s keypresses via dual tone multi-frequency (DTMF) keypresses or voice input via speech recognition. The collected keypresses or voice input is saved to a variable on the call which can be accessed immediately or later on in the flow to control which flow path the user traverses.
If speech recognition is used for input collection, the call is billed for the duration of the speech-to-text (STT) collection in 15 second increments.
*Disclaimer: You will need to have Voice API enabled on your account to do this you will need to open a ticket with support.
Last Updated: 1/21/25
To use the widget simply click and drag it into the builder.
Note that naming the widget can help identify it with other tools later on. It is case sensitive, does not accept spaces or dashes
The Collect Input widget has the following configuration abilities under User Input:
Input
This is how you want to collect the information from the caller, it includes 3 options:
- Digits and Speech
- Digits
- Speech
Language
The language to use for speech recognition
Number of Digits
The number of digits you expect to receive from the caller, leave blank for no limit
Finish on Key(s)
The special key press to end the input, defaults to #
Profanity Filter
Replaces all but the first character of each profane word with a '*' character (e.g., 'f***').
Timeout
Allow with finish on key(s), the timeout is the time in seconds to wait for the caller to enter digits (or speak if no speech timeout is specified).
Speech Timeout
The number of seconds to wait for the caller to speak. This property can either be an integer time in seconds or the word "auto". Setting this property to "auto" causes speech recognition to stop after a pause in speech.
The Collect Input widget has the following configuration abilities under Action Settings:
Action Type
Say Text: allows you to use our built in voices to say text.
Play Audio: allows you to play an audio file.
Say Text Action
Text Language
Select the Language
Voice
Select the voice
Say Text
Enter the Text to Say
Enable Translation (optional)
Displays the options for translation. The text that will be translated is "Say Text"
Language Model (optional)
The type of translation model and vendor to use. Currently we only support Google as a vendor. Text is billed in 100 character intervals and pricing is based on the model selected.
- [Google] Large Language Model: use Google’s LLM to translate text. The LLM supports fewer language pairs than the NMT model, but large language models tend to generate more natural sounding translations.
- [Google] Neural Machine Translation: use Google’s NMT model. The NMT model supports many more language pairs, but may not be as natural sounding as a LLM.
Translate From (required if Language Model is selected)
The language to translate from. The input text should be typed in this language. Automatic language detection is only supported if the user chooses to use the NMT model.
Translate To (required if Language Model is selected)
The language to translate to.
Loop(s)
The number of times to play the audio or TTS.
Play Audio Action
URL
Provide the URL for the audio file location
Loop(s)
The number of times to play the audio or TTS.
Conditions
You can add conditions on the input that is received.
Action (required)
Sets whether to check collected digits or speech. The options are:
- Caller says: checks condition against collected speech
- Caller presses: checks condition against collected digits
- Caller presses or or says: checks condition against collected speech and digits
Our example in the next step takes the speech and digits and uses conditions to route them differently.
*Note that if using a number use the digit. For example, "one" should be "1
Now, you will want to connect the Collect Input widget.
Adding the Collect Input widget is completed!
To continue the flow, connect more widgets to the opposite end of the Collect Input widget:
Now connect more widgets! Click on the links below to view more tutorials:
Voice
Messaging
Tools
Control
