Our great sponsors
-
-
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
-
DeepSpeech
DeepSpeech is an open source embedded (offline, on-device) speech-to-text engine which can run in real time on devices ranging from a Raspberry Pi 4 to high power GPU servers.
-
-
-
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
-
ProjectAlice
Project Alice is a smart voice home assistant that is completely modular and extensible.
-
I've heard of Mycroft before but have no personal experience with it: https://mycroft.ai
It seems like they have an open source client that you can run on your own hardware but it's still dependent on the backend services that Mycroft-the-company provides. Perhaps their privacy stance is more palatable to you, though?
I think there used to be a project that was fully on-device but it got bought and consumed by Sonos
Maybe this is an obvious answer, and therefore not one you're interested in, but there's Siri.
Although audio is sent to the cloud, you can choose whether it is stored or not, and whatever you pick, Apple's privacy policy is very strict.
HomePods are mostly advertised as music playback devices, but I mostly use mine as a HomeKit control device.
Homebridge allows you to control non-HomeKit devices via Siri: https://homebridge.io
This is more of a "DIY" approach but all the tools are there for FOSS and OSHWA solutions.
Mozilla has DeepSpeech [0] and, while not as advanced as the stuff from Google or Amazon, my experimentation left me feeling pretty hopeful that it could reliably recognize at least keywords.
The Raspberry Pi is quite capable though you'll probably need some dedicated microphone to reliably catch voice data. I know ReSpeaker [1] but maybe some off the shelf conference USB microphones would work as well.
[0] https://github.com/mozilla/DeepSpeech
[1] https://wiki.seeedstudio.com/ReSpeaker_4_Mic_Array_for_Raspb...
I've heard of https://getleon.ai before, not sure if it fits your needs. A quick browse through their docs suggests that you'd have to write a package for the lights, but it doesn't seem that hard at first glance.
(It might be harder than it looks especially if your lights' API isn't documented well.)
It's unfortunate that another comment that links to Rhasspy has been downvoted (I assume because it lacked any other context) so I wanted to mention the project with some additional context: https://rhasspy.readthedocs.io/
While I've not used the entire Rhasspy project myself (but trying it out is on the long list of things to do :) ) I have used the offline Text-To-Speech sub-project Larynx...
...and it is amazing!
Larynx is significantly ahead in terms of quality of output & variety of voices (fifty--across multiple languages, accents & genders) of any other FLOSS Text-To-Speech project I've tried.
I think the relative new-ness of the project is part of the reason Larynx (https://github.com/rhasspy/larynx/) currently flies under the radar.
If the rest of Rhasspy is as good as Larynx I'd imagine it's worth trying out.
Larynx demo video: https://www.youtube.com/watch?v=hBmhDf8cl0k
Samples of pre-trained voices: https://rhasspy.github.io/larynx/#en-us
I can vouch for Rhasspy, it's an amazing and flexible piece of software, though it does require some setup and tech knowledge (albeit with a usable web GUI); and it's very DIY on defining the actual voice commands. I recommend pairing it with Node-RED [0] for routing commands to devices, it has plugins for most things.
The only thing I struggled with was getting the wake-word config right: I could never find the right balance point where it responded every time, without also having annoying false positives, so I ended up turning it off. It does support multiple wake-word engines; I'm gonna have another go with Picovoice Porcupine now that they're opened up custom wake-word training for free.
I'm most heavily experienced with Rhasspy's sister project, voice2json [1], which I used to build a voice-controlled car jukebox [2], and it's been working fantastically. (It triggers from a Bluetooth remote, so no wake-word issues.)
For hardware, Raspberry 3/4 perform quite well, and strong recommend for ReSpeaker [3] for audio (either usb or 4-mic hat).
[2] https://github.com/lukifer/voicetunes
[3] https://www.seeedstudio.com/category/Speech-Recognition-c-44...
I can vouch for Rhasspy, it's an amazing and flexible piece of software, though it does require some setup and tech knowledge (albeit with a usable web GUI); and it's very DIY on defining the actual voice commands. I recommend pairing it with Node-RED [0] for routing commands to devices, it has plugins for most things.
The only thing I struggled with was getting the wake-word config right: I could never find the right balance point where it responded every time, without also having annoying false positives, so I ended up turning it off. It does support multiple wake-word engines; I'm gonna have another go with Picovoice Porcupine now that they're opened up custom wake-word training for free.
I'm most heavily experienced with Rhasspy's sister project, voice2json [1], which I used to build a voice-controlled car jukebox [2], and it's been working fantastically. (It triggers from a Bluetooth remote, so no wake-word issues.)
For hardware, Raspberry 3/4 perform quite well, and strong recommend for ReSpeaker [3] for audio (either usb or 4-mic hat).
[2] https://github.com/lukifer/voicetunes
[3] https://www.seeedstudio.com/category/Speech-Recognition-c-44...
I was working on a solution for that a few years ago with an API for raspberry pi
https://github.com/victorqribeiro/rpiapi
Here you can see I'm using voice commands to drive a car
https://github.com/victorqribeiro/raspberryCar
I mean, with a little bit of code and some 3v relays you can achieve what you want
I was working on a solution for that a few years ago with an API for raspberry pi
https://github.com/victorqribeiro/rpiapi
Here you can see I'm using voice commands to drive a car
https://github.com/victorqribeiro/raspberryCar
I mean, with a little bit of code and some 3v relays you can achieve what you want
Consider Project Alice. OSS runs on Raspberry Pi or AMD container. https://github.com/project-alice-assistant/ProjectAlice
The only viable option that I found that could reliably infer commands from speech is https://github.com/Picovoice/rhino
Unfortunately it is not open source (the GitHub just has binary blobs) and requires an account to log in to generate and download model files, but the accuracy is great and you can use it to send commands to Home Assistant to turn lights on/off etc.