[yoyotechKnows] built an Alexa-controlled garage door opener after his Liftmaster stopped working. Now all he has to do is holler at his mobile phone and he can raise and lower his garage doors at will.
His project is based around a Photon WiFi kit, with a pair of LCC 120 digital relays triggering the two doors, reed switches, and a serial-equipped LCD to display door status, with Alexa, IFTTT, and OpenHab to process the commands. You can find his code in the project writeup.
Currently he has a LCD display informing him of the status of each door, hot glued a reed switch to keep track of whether each one is closed. This might seem a little bit extraneous since he can also just look at the doors from within the garage. However, he’s thinking about putting the display inside his house. But couldn’t he just ask Alexa?
We love us our home automation here at Hackaday, with everything from swimming pools to chicken coops rigged for app control and datalogging.
Continue reading “Have Alexa Open Your Garage Door”
The Raspberry Pi is the perfect candidate for Google’s AIY where you can talk to a cardboard box with some electronics in it. [BuddyCasino] took on the challenge of squeezing an Alexa Client in an ESP32 and to make things interesting, a bunny rabbit was chosen as the host of the virtual assistant.
A few months ago, we did a teardown of the Google AIY Kit where [BuddyCasino] commented that he managed to port the Echo Dot client into and ESP32. Sure enough, the video below shows a demonstration of the build in action. The project uses the MAX98357A which is the same I2S DAC used in the Google AIY Voice Hat. For the microphone, the device is again an I2S component however unlike the Google AIY kit which uses the SPH0645LM4H, [BuddyCasino] opted for the ICS-43434.
Two NeoPixels are employed as visual indicators for various purposes. This project is an excellent example of how simple and cheap modern-day designs have become. We are hoping to see the author add more features to the design and who knows maybe we will see a Google Assistant port on the ESP32 in the future. Check out the original teardown for more inspiration. Continue reading “Alexa In A Bunny Rabbit”
It is an inevitability that following swiftly on the heels of the release of a new device there will be an announcement of its rooting, reverse engineering, or other revealing of its hackability. Now the device in question is the Amazon Echo, as MWR Labs announce their work in persuading an Echo to yield the live audio from the microphone and turn the voice assistant device into a covert listening device.
The work hinges on a previous discovery and reverse engineering (PDF) of Amazon’s debug connector on the base of the Echo, which exposes both an SD card interface and a serial terminal. Following that work, they were able to gain root access to the device, analyze the structure of the audio buffers and how the different Echo processes use them, and run Amazon’s own “shmbuf_tool” application to pipe raw audio data to a network stream. Astoundingly this could be done without compromising the normal operation of the device.
It should be stressed, that this is an exploit that requires physical access to the device and a bit of knowledge to perform. But it’s not inconceivable that it could be made into a near-automated process requiring only a device with a set of pogo pins to be mated with an Echo that has had its cover quickly removed.
That said, inevitably there will be enough unused Echos floating around before too long that their rootability will make them useful to people in our community. We look forward to what interesting projects people come up with using rooted Echos.
This isn’t the first time we’ve covered the use of an Echo as a listening device.
Via Hacker News.
Amazon Echo image: FASTILY [CC BY-SA 4.0].
We’ve all probably done it — gazed up at a passing jetliner and wondered where it was going and what adventures its passengers were embarked upon. While the latter is hard to answer, the former just got a bit easier: just ask Alexa what the plane is.
Granted, [Nick Sypteras]’s Echo Dot isn’t quite omniscient enough to know exactly what plane you’re looking at. His system benefits from the constraints offered by the window of his Boston apartment — from the video below, we’d guess somewhere in Beacon Hill or the West End — that offers a view of the approach to Logan Airport. An RTL-SDR dongle receives the ADS-B transmissions from all aircraft in the vicinity, and a Raspberry Pi does a lookup, picks the closest plane, and scrapes the departure and arrival airports from FlightRadar24. Alexa does the rest, but we have to confess that hearing “Boeing seven hundred eighty-seven” rather than “seven eighty-seven” would drive us nuts.
If you don’t have the limited view of an airport approach that makes [Nick]’s hack workable, maybe a plane-spotting robot camera would work better for you.
Continue reading ““Alexa, what plane is that?””
The Hippopotamus is the most dangerous large animal in Africa. The Internet of Things will kill us all. What do you get when you combine the two? Hungry hungry. [Mike] took the classic game Hungry Hungry Hippos and turned it into an amazing and amusing Internet of Things device with voice recognition and machine vision.
Hungry Hungry Hippos is a child’s (board?) game designed to teach children the virtue of gluttony. The board is surrounded by four lever-actuated plastic hippopotami, and the object of the game is to mash a lever and collect marbles in the mouths of these piggish pachyderms. [Mike] automated this game with four servos connected to these levers, with each servo controlled by a W65C265SXB single board computer. Yes, this project has code written in 6502 assembly.
Taking this a step further, [Mike] is using a Playstation 3 camera connected to a netbook for image processing. When the camera detects a marble in front of a particular hippo, that hippo becomes hungry hungry. Autoplaying Hungry Hungry Hippos. What a fantastic time to be alive.
The Internet of things connectivity? [Mike] also made these hippos controllable via Amazon’s Alexa with the help of an Electric Imp. To activate the blue hippo, all [Mike] needs to say is, “Alexa, tell game that blue hippo needs to eat.” It’s an Internet of Things computer vision AI hippopotamus. We all knew technology was eating us alive, but we never thought technology was hungry hungry.
You can check out [Mike]’s demo videos below.
Continue reading “Internet of Hungry Hungry Things”
Back in May, Amazon announced the Echo Show, its new version of Alexa with a 7 inch touchscreen. The Echo Show is an interesting device, but will the great unwashed masses pony up $229 to buy the show? That’s $50 more than the original Echo, or $180 more than the Echo Dot. With 5.2 million units sold in 2016, Echo has been a resounding success. This has been in part due to Amazon’s open approach to the API. Anyone can build an Alexa compatible device using a Raspberry Pi. Google has (finally) followed suit with their Home device.
It’s not just the hardware that is accessible. Skills Kit, the programmer interface for extending Echo’s functionality, is also open. At CES this year, Alexa was the belle of the ball. Third party devices are being introduced from all corners, all of them connecting to Amazon’s cloud and responding to the “Alexa” keyword.
The Echo Show takes the family in a new direction. Adding a touch screen gives the user a window on the the world not available with voice interactions. Echo Show also includes a camera, which opens up a whole new set of privacy and security questions. Amazon touts it as a device for viewing security cameras, watching YouTube videos, and making video calls. This puts Echo Show dangerously close to the internet appliance category, essentially a barren wasteland littered with the corpses of previous devices. Does anyone remember when Palm tried this with the 3Com Ergo Audrey? How about the i-Opener? Will Alexa persevere and succeed where others have failed? A lot of it will depend on the third party developers, and how Amazon treats them.
Continue reading “Amazon Echo Show”
Nothing makes us feel more like we’re on Star Trek then saying “Computer, turn on desk light,” and watching the light turn on. Of course, normal people would have left the wake up word as “Alexa,” but we like “Computer” even if it does make it hard to watch Star Trek episodes without the home automation going crazy.
There’s a lot of hype right now about how voice recognition and artificial intelligence (AI) are transforming everything. We’ve even seen a few high-profile types warning that AI is going to come alive and put us in the matrix or something. That gets a lot of press, but we’re not sure we are even close to that, yet. Alexa and Google’s similar offerings are cool, there’s no doubt about it. The speech recognition is pretty good, although far from perfect. But the AI is really far off still.
Today’s devices utilize two rather rudimentary parts to provide an interaction with users. The first is how the devices pattern match language; it isn’t all that sophisticated. The other is the trivial nature of many of the apps, or — as Alexa calls them — skills. There are some good ones to be sure, but for every one useful application of the technology, there’s a dozen that are just text-to-speech of an RSS feed. Looking through the skills available we were amused at how many different offerings convert resistor color codes back and forth to values.
There was a time when building electronics meant learning the resistor color code. With today’s emphasis on surface mount components, though, it is less useful than it used to be. Still, like flossing, you really ought to do it. However, if you have an Amazon Alexa, it can learn the color code for you thanks to [Dennis Mantz].
Don’t have an Alexa? You can still try it in your browser, as we will show you shortly. There are at least eight similar skills out there like this one from [Steve Jernigan] or [Andrew Bergstrom’s] Resistor Reader.
Continue reading “Alexa, Sudo Read My Resistor! A Challenge for Hackers”