Personal mycroft backend alternative to mycroft.home, written in flask
you can run it, but why would you before it's finished?
Pull Requests welcome
Mycroft Backend API docs can be found here
from source
git clone https://github.com/MycroftAI/personal-backend
cd personal-backend
pip install .
from pip
pip install git+https://github.com/MycroftAI/personal-backend.git
configure backend by editing/creating ~/.mycroft/personal_backend/personal_backend.conf
{
"backend_port": 6712,
"website_port": 5000,
"ssl": true,
"ssl_key": "/home/user/.mycroft/personal_backend/certs/MycroftPersonalServer.key",
"ssl_cert": "/home/user/.mycroft/personal_backend/certs/MycroftPersonalServer.crt",
"mail_port": 465,
"mail_server": "smtp.googlemail.com",
"mail_user": "[email protected]",
"mail_password": "xxx",
"secret_key": "MY_PRECIOUS_SECRET_KEY",
"salt": "MY_TABLE_SALT"
}
change url in "server" section in your default mycroft config
// Address of the REMOTE server
// Override: none
"server": {
"url": "https://0.0.0.0:6712",
"version": "v0.1",
"update": true,
"metrics": true
},
start backend
from personal_mycroft_backend.backend import start_backend
start_backend()
start frontend
from personal_mycroft_backend.frontend import start_frontend
start_frontend()
more examples here
-
get location
-
geoip location default
-
get config
-
get device settings
-
patch device settings
-
pairing process
-
send mail
-
multiple STT engines supported (google, wit, ibm, kaldi, bing, houndify, govivace, deepspeech)
-
sql database
-
user email confirmation
-
local browserless GUI if running on same machine
-
website (status: WIP)
-
store metrics (WIP - integrating into frontend, sql db and api endpoints functional)
-
most of the frontend, are you a web dev? team up with me!
-
STT json settings in db (for google cloud stt engine)
-
skill settings
-
wakeword upload
-
wolfram alpha api
-
weather api
-
local data storage (audio recordings)
-
if opt in proxy data to mycroft (utterances/hotwords/metrics)
-
everything else i forgot
-
new functionality
- mimic 1 remote TTS
- user voice print from uploaded utterances / wakewords
- train precise/snowboy models
Thanks goes out to JarbasAI for creating this initial implementation. You can find the archived original repo here.