Community Blog Buy Now

Misty Community Forum

Having Misty record a response log of a conversation?

We are trying to find a way to have Misty keep a log of a conversation they have with a user. With responses, we want Misty to keep a log of response through a user that she recognizes (via facial recognition she has). Is there anyway we can have Misty record a log of each conversation she has, and record the responses of each user input vocally?

@sollazzo08 is working with me on this as well so I will add him to the thread.

Hey @kywelch17, great question. This is definitely doable, and there are a number of possible approaches. It’s just a matter of how you prefer to implement things :slight_smile:

Without knowing your requirements or setup, I think you’ll probably want to consider using a third-party natural language processing service to handle most of the voice interactions. Misty can capture speech, send the audio file to an API, and process the response from whatever service you want to use. This might be how you would get the text content of the user utterance back into your skill. You can then do whatever you want with that data – you can save it to the long-term storage that’s associated with that particular skill (the command for this is misty.Set() in JavaScript, and you’ll soon be able to do this with Misty’s .NET SDK), or you could have her send it off to some database that you’ve set up on a remote device. If you’re already using C#, you can even write the text to a file on Misty’s 410 processor.

And that’s just one possibility – don’t let this suggestion limit your thinking! Feel free to ask more questions or share more information about your use case here.

If you haven’t already seen this, the Misty-Concierge-Template is a great way to learn how to use Dialogflow (Google’s NLP service) with Misty’s skills.

I try to run the skill, it worked a few times however whenever now it won’t recognize us saying, “Hey Misty.”

Did you examine the logs? (For example, you can get these via the Command Center.)

Yeah, we didn’t realize there was a timeout on the hearing start phrase function, our bad on that end. We will definitely look more into the repo John sent us!