All the features at a fraction of the cost.
Tooling through instagram feeds, and due to the people we were connected to in Chicago, we seemed to be presented with the occasional advertisement for Glance. Glance is a smart mirror product designed and produced by some local friends and in a far more mature company structure than anything that we’d been pursuing with Double Bear Rolled. The price point to acquire a Glance product was well outside of the means for our learning-intensive budgets. Plus the Glance Connect feature - a robust software system designed to empower control over what’s displayed on a Glance product - we couldn’t compete with or really had the background or core product to justify building ourselves.
So we knew we wanted to experience a smart mirror, but didn’t have the budget or expertise to build a fully fledge product. Queue up the DIY articles and pinterest boards: We’re going to make one.
We had a tri-fold bathroom mirror at our disposal that was somewhat modular. We had a few (older) Nexus tablets lying around. And we needed an excuse to start buying power tools (small ones, at least). Our intention was to follow an Adafruit tutorial that directly nailed the concept we wanted to explore. Through some googling we found a local (to Illinois) supplier of custom-cut one-way plexiglass (“glass”) and found a starting point of building software in an Android app from Hannah Mitt.
Since this effort was more focused on building something tangible than exclusively building software, we needed to ensure we had the right tools available. Working with a 1970s-era vanity mirror meant we had to expect the materials to be worn and prone to breaking. (We also didn’t get permission from anyone to make such a replacement in the mirror, so anyone that might wag a finger at this effort was going to need some assurances we could undo the work we were doing.) Measuring the mouth of the mirror frame to ensuring we had the right hole on the back of the one-way “glass” was critical to this project’s success. Ultimately if we wanted the illusion of a smart mirror, we needed to ensure that illusion wasn’t broken by back lighting or cables (slowly, over time) separating the tablet from being flush with the glass.
We also didn’t anticipate or account for an older Android tablet having any kind of controls. So while the rest of the apartment was wired up with Google Home devices, yelling “hey Google” to change lighting effects in different rooms, meant that same yelling throughout the house was going to affect the tablet’s ability to trigger the Google Search Assistant. Also, with the tablet flush against the “glass” we wouldn’t have touch control through the mirror’s “glass” to adjusted, edit, or manipulate the tablet. Whatever we were going to setup to be displayed had to be correct by the time we applied the tablet to the back of the glass.
We solved a lot of these problems by sticking to the mentality that we’re creating the illusion of a smart mirror, and not building some fancy marketable product. What’s not pictured, nor will we admit if questioned later, is that we held the tablet in place with painter’s tape. While the backing of the “glass” and the “glass” itself are glued, stabled, and adhered to the tri-fold mirror’s frame, the tablet itself is pretty easy to take off the back of the whole setup. (This came in handy later, as the older tablet would randomly crash from time to time and the display would need to be turned back on, which required engaging with the touch screen, with the tablet removed from the mirror.)
Utilizing a modified version of the Adafruit tutorial and slightly tweaking Hannah Mitt’s original Android app, we landed a smart mirror experience that worked for our needs. One of the cool features built into the android app was a small scale facial recognition library that used the front facing camera to detect if you were smiling or not. If the probability of you smiling was low, strings of text (like the pictured “Cheer up, bud”) was presented on the mirror. And if you’re already smiling you get other encouraging prompts (like the pictured “Looking flyyy”) that really made this project a gas.
A few weeks after pursuing this project, coming back to the mirror was helpful in that it gathered events from the calendar, had the encouraging facial recognition greatness, but most importantly, provided timely information in a hands-free environment. Maybe not everyone requires both hands when brushing their teeth, but having the smart mirror scroll through the day’s top news headlines or show the next upcoming appointments on a personal or work calendar saved a few minutes and a few steps in the morning routine. While we haven’t tracked any data or really quantified the amount of time saved, just riffing on the monetary value of this project: saving a few minutes each weekday (let’s say 3 minutes) adds up over the course of a year. For a year with 260 workdays, those 3-minutes-per day add up to around 13hrs per year saved in morning routines. Sure it’s a year, but for a device and experience that puts the user in a better mood each morning, informs them in a passive and proactive way of what their day will entail is providing value.
And let’s talk sheer numbers, for a person who’s time is worth $20/hour, saving 13 hours per year is about $250. The “glass” cost about $30, the tablet was $20 (came in a two-pack when purchased), and other odd materials cost under $20 (glue, Dremel bits, sand paper, etc) - so not including our time (the cost of which is offset by learning) the materials for the project cost under $100. Easily a profitable project if going by these numbers. As of this writing, it’s been 2 years using this smart mirror and it still provides content and value on a daily basis.