Calling via Mycroft

Skill name: Call-Mom-skill

User story: If I understand the adds for Alexa correctly, it is able to perform calls, I guess via a connected phone, or maybe VoIP from the landline (?). It would be neat if the Mark I or II could do the same. Or my home PC, while ad it.

As a person who calling a lot while doing other stuff, real hands free calling is very handy. Some headsets offer own software for that, unfortunately some (mine for example) switched to Alexa. While having mycroft at home anyway, it would be a perfect way to use the phone while still having the hands perfectly free to continue cooking, woodwork, etc.

What third party services, data sets or platforms will the Skill interact with? Mobilephone, routers maybe?

Are there similar Mycroft Skills already? no

See https://github.com/MycroftAI/mycroft-skills for a list. If so, how could they be combined? no

What will the user Speak to trigger the Skill? hey mycroft, call …

For example “What’s the weather like?” or “How much is the Canadian dollar worth right now?” or “Play Aussie Rock music”

What phrases will Mycroft Speak?

For example, “The weather is cloudy with a top of 60F”, “The Canadian dollar is worth 90 US cents”, “Now playing Aussie Rock on Spotify”

What Skill Settings will this Skill need to store?

See https://mycroft.ai/documentation/skills/skill-settings/ for more information

Other comments?

_It will also be usefull to have it for future android use. _

Hey roever, thanks for the great suggestion.

I wanted to flag that there’s work underway on an Intercom Skill which is currently just between Mycroft devices on the same LAN, however I know the developer is keen to work toward external voice calling in the future. This is definitely a big undertaking so if other developers are keen to support that effort please get in touch.

3 Likes

Hi,
I’m wondering if calling via Mycroft (e.g. over the bluetooth connection with a phone) is on the map? It would be really cool if you just could say “call mom” and then call over microphone and speaker of Mycroft with your mom.
Best,
JulianFV

Hey there, I doubt our internal team will look at this until we’re releasing an Android application. However there are some community efforts bringing Mycroft to platforms like the PinePhone and Librem 5 so I’m sure there are plans for calling functionality for those at some point.

I am the aforementioned PinePhone effort person, and I won’t put an ETA on the actual phone call. But I will say that dialing the phone is my reference implementation for, “Make the device do things over a thin client,” and that we know exactly how it’s gonna work. Just a bunch of code to be added to an existing community project.

A whole bunch.

Isn’t a VOIP account and the VOIP Mycroft skill from @JarbasAl not a lot easier and already available?

Yes, but the overwhelming majority of people would rather use their phone, and plenty of others are rationally unwilling to pay for an extra phone service when they already have one.

Plus, what kind of mobile integration would we be making if we didn’t include bog-standard voice-control features that literally predate Siri?

I wonder if it would be that complicated to get a connection over the bluetooth adapter of the pi to a phone and then to start a call? I’m a new one in this field but wouldn’t that be possible?

I am making some progress, FWIW, with HiveMind’s take. The biggest obstacle right now is that I have no cell signal at home, but my hope is to have my PinePhone attempting to place a call in the next few days, at which point I’ll take the phone down to where I get service and tunnel home, give it a shot. Already know the thing works over Wireguard.

This does not, however, help Mycroft users who don’t run HiveMind, and it doesn’t get as far as “Use this smart speaker as the I/O device for the phone’s current call.”

Frankly, the number of steps required to make that happen is long enough that I don’t really want to type out why it’s a much bigger ask than it sounds. Tl;dr you’d speak the utterance, wait for your particular TTS to parse it, then the intent parsers are pretty fast. You need a skill that triggers a “dial phone” type deal, switches the device’s bluetooth situation around so that it can serve as the phone’s “headset”, opens the connection, and sends the “place call” instruction, all in a user-tolerable amount of time.

And for what?