百科.dev
全部条目AI 编程趋势榜开源项目技术资讯提交条目
登录
< 返回工具列表
M

MMM-SUSI-AI

> 编程语言
开源

Magic Mirror 的 SUSI.AI 实现

1.5K stars0 点赞0 次浏览
访问官网GitHub

工具介绍

Magic Mirror 的 SUSI.AI 实现

Susi Magic Mirror Module

SUSI.AI Module for Magic Mirror Project

This project aims at creating a Magic Mirror Module to provide intelligent answers via Susi directly on your Magic Mirror.

Magic Mirror Project: https://github.com/MichMich/MagicMirror

Current Status

Currently, magic mirror module for Susi is in working state. You can invoke Susi via hotword 'Susi' and ask any query, Susi will reply back with answer. Notably,

  • Hotword Detection is working via Snowboy Hotword Detection
  • Speech Recognition is Working via Microsoft Cognitive Services: Bing Speech API
  • Text to Speech is working via Bing Text to Speech.
  • Answer and Map Action Type by Susi API Working.
  • Authentication with Face Recognition with MMM-Facial-Recognition module.

Next Steps

  • Add better visualization support in module
  • Display data by various action types on Mirror Screen.

How to install?

Hardware Requirements

  • A Raspberry Pi or similar development board.
  • A screen with 2 way mirror.
  • A USB Microphone.
  • A Camera for Face Recognition (Optional). If you are using a webcam it has Microphone built in , so you do not need USB Microphone.
  • Keyboard and Mouse. It is optional but recommended for initial setup. You may also SSH into your development board.

Software Requirements

  • NodeJS (Recommended version 6+). Raspbian repositories have old NodeJS. Install using nvm
  • Node Package Manager (npm)
  • Some additional dependencies. Install using
sudo apt install sox libsox-fmt-all libatlas-dev libatlas-base-dev

Install Magic Mirror

  • git clone https://github.com/MichMich/MagicMirror
  • cd MagicMirror
  • npm install & npm start
  • If you wish to start in development mode, run npm start dev

This will start Magic Mirror with basic modules on your screen.

Install Susi Module

  • Go to modules directory in Magic Mirror project folder and clone Susi MagicMirror repository

git clone https://github.com/fossasia/MMM-SUSI-AI.git

  • Go to susi_magic mirror directory, run npm install
  • Also run, npm run electron-rebuild
  • Run ./install-flite.sh to install Flite TTS Voices.

Add Susi Module to MagicMirror

  • Edit config/config.js in the MagicMirror directory
  • Add following to modules JSON array to enable Susi Module,
{
	module: "MMM-SUSI-AI",
	position: "top_bar",
	config: {
	    hotword: "Susi"
	}
}

or if you want to use SUSI in authenticated mode with Face Recognition

  • Install MMM-Facial-Recognition Module to your MagicMirror by following instructions on the offical Github page
  • Train Face Recognition for 1 or more person. You need to have a SUSI account for each one of them. Sign up at https://accounts.susi.ai
  • In the config file (config.js), add the following lines. If there are multiple users, add more user object to users array.
{
	module: "MMM-SUSI-AI",
	position: "top_bar",
	config: {
	    hotword: "Susi",
	    users: [
	    {
	        face_recognition_username: "NAME_FOR_THIS_USER_USED_IN_MMM-Facial-Recognition"
	        email: "YOUR_EMAIL",
	        password: "YOUR_SUSI_PASSWORD"
	    }
	    ]
	},
	classes: "default everyone"
}
  • Save the config.
  • Start the MagicMirror again by npm start

Now, you can invoke Susi via "Susi" Hotword. Once Hotword Detected, your MagicMirror screen will blur. Ask you query by voice after that, Susi will speak back the reply and Mirror Screen will go in original state.

Additional/Optional Setup

If you want better experience with SUSI AI on your MagicMirror, disable the default 'Compliments' module. To disable the 'Compliments Module', comment or remove the following lines in configuration file in config/config.js

{
			module: "compliments",
			position: "lower_third"
},

Got any problems ?

While running if you get a log statement mentioning SnowBoy module not found, run npm run electron-rebuild

Project is still in alpha, please report an issue if you experience any problems with logs/screenshots.

Credits

The code structure of this module has been inspired and taken from Awesome Alexa Module for Magic Mirror.

GitHub Issues· 7 开放

在 GitHub 查看全部
  • #42

    What is the status of this?

    更新于 2018年5月12日
  • #40

    npm install throws error

    更新于 2018年2月14日
  • #41

    Getting this error......what is it?

    更新于 2018年1月5日
  • #32

    Use React to create UI

    更新于 2017年9月24日
  • #31

    Add Google Text to Speech

    更新于 2017年7月21日
  • #30

    Add Live Speech Recognition support for intermediate results

    更新于 2017年7月17日
  • #22

    Process SUSI Response for all different Action Types.

    更新于 2017年7月12日

核心特点

  • •Hotword Detection is working via Snowboy Hotword Detection
  • •Speech Recognition is Working via Microsoft Cognitive Services: Bing Speech API
  • •Text to Speech is working via Bing Text to Speech.
  • •Answer and Map Action Type by Susi API Working.
  • •Authentication with Face Recognition with MMM-Facial-Recognition module.
  • •Add better visualization support in module
  • •Display data by various action types on Mirror Screen.
  • •A Raspberry Pi or similar development board.
  • •A screen with 2 way mirror.
  • •A USB Microphone.

> 标签

TypeScript

暂无评论,来聊聊你的看法吧

> 工具信息

发布日期2026年8月1日
最后更新2026年9月17日
分类编程语言
定价开源

> 相关工具

T
TypeScript
JavaScript 的超集,为前端与全栈提供静态类型
P
Python
通用编程语言,广泛用于 Web、数据与 AI
G
Go
Google 推出的简洁高效系统语言