How to Install alwaysAI on a Mac
by Taiga Ishida | Apr 30, 2020 | Computer Vision | 5 minute read
The process of developing computer vision applications has been greatly simplified by alwaysAI, which now includes native support for Mac OSX (Mojave and Catalina), and enables developers to get started prototyping applications right away with very little setup required. By the way, you can also learn how to install alwaysAI on Ubuntu.
Using the convenient starter apps provided by alwaysAI allows developers to build computer vision applications such as object detection on a local machine (which can be easily deployed on any edge device using the alwaysAI CLI) so that less time is spent sourcing and provisioning devices upfront, and more time is spent creating applications and making a real impact as quickly as possible.
If you haven't already, you can create your alwaysAI account here.
Here’s how to install the alwaysAI environment on your Mac:
1. From the alwaysAI homepage, navigate to the 'Docs' link next to the 'Dashboard' button on the upper right.
2. Click on Get Started, either at the top of the menu on the left-hand side, or on the upper-left tile on the landing page.
The alwaysAI Mac installer is not signed by Apple and will prompt an “unidentified developer” warning. Follow the directions below to install it.
Once the installer is downloaded, Mac users should ctrl-click to show the package in their finder:
Then ctrl-click the package file and choose Open:
From here you will get the following security warning. Choose Open to install:
After opening the package, you will be taken through install steps as follows:
The install summary screen includes important instructions for users who have previously installed the alwaysAI CLI via NPM:
At this point it is a good idea to open a terminal window and type aai -v to verify the alwaysAI CLI is installed and available. At any time you can check CLI options and syntax by typing aai. Most of the CLI commands are interactive as well.
And with that, you’re ready to go!
Navigate to the dashboard and click 'create new project'. Then select 'start from scratch' and type in the project name 'hello_world'.
Then click 'next' and on the subsequent page click 'create new project' (skipping adding a model for now).
Next, download the starter apps. Use your terminal to navigate to a directory where you would like the start applications installed. Then type into the command line aai get-starter-apps. The starter applications will be downloaded into a subdirectory (always-starter-apps):
This directory will include not just Hello World, but many other starter applications as well. Real-time applications will use either your Mac’s built-in camera or a webcam by default. All others, including Hello World, take a set of images that will have been downloaded with the starter application. See the alwaysAI docs for instructions on changing data sources.
Then navigate to the hello_world directory in your terminal and enter into the terminal: aai app configure . Then select the 'hello_world' project and choose 'local computer' as the device.
Run aai app install at the command line; then you should see:
Finally, run aai app start to execute the application:
The CLI will give you a URL (highlighted in the image above) where you can see the application running. Just paste it into a browser window to invoke the alwaysAI streamer:
That will conclude your Mac set up!
For additional setup information and help, return to the Development Computer Setup page for more details on computer set up.
RELATED: Learn how to create and run an object detector starter app in minutes
