SOLVED help with a tricky dial plan similar to Lenny, if he married Wolfram Alpha with Googles permission

[ABSGINC]

Trying to marry some examples of Lenny and SMS Blaster and Wolfram Alpha, to vision a better voice automated response I'm dreaming up..

A handful of you in this group probably have seen the dial plan for Lenny, the telemarketers worse nightmare.

I like how he uses the talk detect and waits for someone to stop speaking before he goes on to his next predetermined prompt. I am trying to accomplish the same thing, in a modified version that after the speaking of Lenny for instance is done, while waiting for silence, i would like to save to a temporary file only the words spoken once waiting for speech and silence executes, so only recording the speaking up till the silence, at that point I would imagine playing a background noise (comcast ivr mimics a computer typing sound) and after I'm done with the logic running, run to the next sub and repeat. Very similar to Lenny.

My first step in accomplishing my goals is figuring out how to record segments of a call throughout the dialplan determined by speaking, then I'll move on to processing the temporary files similar to other examples in incrediblepbx feature sets.

Most call centers now allow you to give a spoken response in addition to pressing DTMF tones, as I've ran across several that prompt only for you to say 'this or that' and no longer have options for press 1 or press 2.

I could hack together something based on how sms blaster or wolfram works, although I am wishing to prevent a user from having to press pound when they are done speaking.

Any Guidance, Much Appreciated!

 

[ABSGINC]

Nothing? -- it's OK, i asked for help too soon. But for anyone else who finds this post searching I owe this reply.

As it turns out the speech recognition agi file available from Lefteris Zafiris' git Hub, also included in the IncrediblePBX feature set, will handle nearly all of what I complained about above. I was unaware of the NOBEEP option or of its ability to detect silence or set timeouts for maximum recording. That beep just made me always think I had to press # too, i never would wait 2 seconds

excerpt from speech-recog.agi

....
# agi(speech-recog.agi,[lang],[timeout],[intkey],[NOBEEP])
# Records from the current channel until 2 seconds of silence are detected
# (this can be set by the user by the 'timeout' argument, -1 for no timeout) or the
# interrupt key (# by default) is pressed. If NOBEEP is set, no beep sound is played
# back to the user to indicate the start of the recording.
# The recorded sound is send over to googles speech recognition service and the
# returned text string is assigned as the value of the channel variable 'utterance'.
# The scripts sets the following channel variables:
# utterance  : The generated text string.
# confidence : A value between 0 and 1 indicating how 'confident' the recognition engine
#     feels about the result. Values bigger than 0.95 usually mean that the
#     resulted text is correct.

In setting a maximum timeout close to the expected spoken response I'm able to set a fail-safe on if there is background noise that could give the silence detection a hard time. This hasn't been an issue though it seems fairly accurate.

In expanding from here, I've been working on a Facebook Messenger plugin of sorts that run on wazo and sends messages with call alerts, but in it's bot nature allows commands to be sent. So away from home on my Facebook Messenger App, possibilities of continued coding, sending a Message to the bot, ask for missed calls, voicemails, queue status, log in or out of queues, schedule a reminder, and even pass an audio file through messenger, the api returns with the message a url of the wav file, and the facebook api allows for a very friendly api in sending back formatted lists and quick reply buttons.

Check the Full Facebook API Documentation for examples and resources ,

and my recent post on the Working Facebook Messenger Alert Example on Wazo 17.12

If anyone would be interested in working with me on some of this and getting something clean and neat as far as a starter package on github please PM me