However, someone here on HN recently posted the open source voice assistant project mycroft:
Fun to mess around with on Raspi and eventually I may link it up with Home Asssistant.
However, someone here on HN recently posted the open source voice assistant project mycroft:
Fun to mess around with on Raspi and eventually I may link it up with Home Asssistant.
https://medium.com/snips-ai/snips-nlu-is-an-open-source-priv...
The whole platform runs on-device which makes it ideal for privacy, cost, and to allow it to run when there is no network, from hotword to ASR and NLU
We are available in English, French, German, and soon Japanese and Korean, with more European languages coming!
Take a look at what some people have built with it: https://github.com/snipsco/awesome-snips
and a few tutorials to get you started: https://medium.com/snips-ai/building-a-voice-controlled-home...
I like Spotify but I’d also like to be able play my local music with voice commands.
On first glance, hmmm: why does this open source AI require that I create an account on their webpage before trying it out?
> Mycroft AI, Inc. maintains a device and account management system known as Mycroft Home. Developers may sign up at: https://home.mycroft.ai
By default, mycroft-core is configured to use Home. By saying "Hey Mycroft, pair my device" (or any other request verbal request) you will be informed that your device needs to be paired. Mycroft will speak a 6-digit code which you can entered into the pairing page within the Mycroft Home site.
[...]
> If you do not wish to use the Mycroft Home service, you may insert your own API keys into the configuration files listed below in configuration.
Bingo I'm in! Thanks a lot. This looks awesome.
First, there is no premium version. The whole platform is free to use. Forever. And we are starting to open-source it.
For the trigger words, I think there is a slight confusion about what the are.
There are actually two kinds of trigger words:
- ones which are universal, and are very hard to train (it takes a lot of data and use NN models). Those are paid because we need to get the data (this is the only reason)
- ones which are much simpler and can be trained on your device. This will come in the next release of the platform!
Stay tuned for the next release where the personal hotword is coming!
Looks like it can all be done offline + OSS except for Speech-To-Text. But I think the key idea is even if it's using cloud for processing, you can at least trust it isn't listening when it shouldn't be (as long as always-on, wake word detection is offline).
You can then use any language to create skills to make it perform actions.
Take a look at how to build a connected speaker with Snips: https://medium.com/snips-ai/building-a-voice-controlled-home...
Which one? And why would you not consider Facebook's speaker but are apparently using somebody else's same product.
https://learn.adafruit.com/using-an-infrared-library/hardwar...
I do something similar, I have an ESP8266 microcontroller hooked up to an IR sender/receiver that I can use to send and receive codes from Home Assistant to control some stuff. The module I use takes care of encoding and decoding the IR and simply dumps it to serial, which the micro reads and forwards to Home Assistant via MQTT (and vice versa). It's this one
https://www.aliexpress.com/item/5V-IR-Infrared-Remote-Decode...
but it's not hard to do it with a bare IR emitter and receiver. Unfortunately I haven't had have much time lately to document many of my projects for other people, but you can learn a lot from Youtube and Google these days.
I use an ultrasonic sensor to detect when the door is open or closed. It's very reliable.
I even wired up Home Assistant to contact me (over sms by sending an email to my number@carrier) if the door has been open for >5 minutes, because I was precisely in your position of forgetting to close it.
With a little z-wave door sensor (just a magnetic reed switch), you could have Home Assistant run an automation that says "If gate is open at 9:30am, close it." so then you don't even have to log in to click it.
[1] https://partofthething.com/thoughts/adding-a-ir-send-and-rec...
Some have a setting to close a few minutes after being opened. Might be simplest.
I've been wanting to add this to my setup. Do you have any more details on what you have done? My main hurdles would be choosing decent cameras that can be used by Home Assistant... Cameras that don't want to phone home all the damn time.
What camera?! Seriously though, I don't necessarily want to write device drivers for a camera, maybe it's time I start looking at usb webcams to use?
There are also a ton of different 3D-printable mounts and jigs for them on sites like thingiverse and instructables.
My favorite[1] comes as a kit with the jig, 3 servo motors, and an add-on board for a raspi. Using this I have a live-feed camera that can rotate up to 180 degress on 3 axis that I can control from anywhere in the world using my phone as long as my internet connection is up at home. All hosted locally, even the live feed frontend. It's not hard to build or set up either, took me a lazy day and I'm far from a good coder.
1: https://shop.pimoroni.com/collections/raspberry-pi/products/...
I remembered incorrectly, that PiMoroni kit does not include the 3 servo motors, you need to buy those separately along with the RPi camera module. Luckily the servos required are super common little blue servos that are all over eBay and such for dirt cheap.
[1] https://partofthething.com/thoughts/raspberry-pi-home-assist...
[2] https://www.home-assistant.io/components/camera.amcrest/