Allows the definition of individual terms, with return values, to recognize voice prompts in a voice recognition session (see <voice_rec>).

Each term must be separated from its return value by a colon; each set of terms and return values must be separated by a comma.

Optional and alternate terms may be defined as follows:

[[the] color|colour] red|crimson:red

In the above example, each of the following terms will return "red":

the color red
the colour red
the color crimson
the colour crimson
color red
colour red
color crimson
colour crimson

In addition, the expression $GARBAGE may be used to match on and discard any extraneous terms:

[[the] color|colour] red|crimson $GARBAGE:red

In the above example, the following input:

the color red is my favorite color

will return:


The return value is placed in the long_poll_value variable in the long polling session that captures the voice input (see <long_poll connect="">).

Multiple <dynamic_grammar> tags may be used within a single <voice_rec> tag.

Child of: <voice_rec>
Parent of: Nothing


<dynamic_grammar>PTN_1:RET_VAL_1, PTN_2:RET_VAL_2</dynamic_grammar>