Does this project spark your interest?

Become a member to follow this project and don't miss any updates

Project Jarvis - A.I Home Automation & Assistant

Intelligent home automation hardware and software on a budget which utilizes a full home personal digital assistant and electricity savings

Similar projects worth following

This project was created on 05/22/2014 and last updated 6 months ago.

We have smart cars, smart phones and even smart T.V's. It's time to bring smart homes to the future.

Project Jarvis is designed with Artificial Intelligence (A.I) in both software and hardware which allows the computer system to make smart decisions based on environmental factors. This means the system can help save energy in your home, reducing your monthly bills exponentially and saving the Earth! That's not all though, Jarvis is a personal assistant found throughout your home and Android device. He can make you coffee in the morning, wake you up with weather and you can even speak to him. You can even tell him to switch off all the lights in your room before you go to bed and it will be done instantly. Do you have a complex maths equation or assignment? Not to worry, Jarvis can answer all your questions and even complete full research reports for you.

It's all driven by sophisticated hardware and software to help make your life, and the life of others, much better.

Key Features (At a Glance)

- A.I Home Automation

- Energy Savings Efficiency

- Voice Recognition and Speech Synthesis

- Enhanced Security and Safety

- Entire House Wide Virtual Assistant

- Sleep Pattern Tracking

- World Wide Web Enabled Home Automation

- Custom Android Application and Computer Software

- Low Cost (Under $200)

- Research Report Compilation, Question Answering (Including mathematical equations)

- Any hacked appliance or device can work with this system

Jarvis In-Depth

Intelligent Automation

Jarvis is the name of Iron Man's computer system which manages his home, hence the name of this project. The main aspect of Project Jarvis is to help save electricity in your home which then helps save the Earth and of course, helps save a bit more money each month. The electricity usage is logged per room into an SD card on an open source micro controller such as an Arduino or Raspberry Pi, I chose Arduino. The code on the MCU then interacts with a larger more complex network known as a neural network which has been programmed in a separate application. This software then makes intelligent decisions using the environment and its sensors to help beat the current months data which is still on the SD card. Jarvis will attempt to save more energy than the previous month without getting in the way of your usual living style.

Who is Jarvis?

Jarvis is a virtual assistant, similar to Siri but with a lot more power. Jarvis is the brains behind all operations, he makes the choices and decisions for home automation and much more. Jarvis is based off Artificial Intelligence and is capable of answering complex questions, maths equations, reading social network messages, conversing, being a top class assistant and much more. Jarvis speaks back through speakers located in the rooms and you can speak to him! That's right, if you have a question, simply ask it wherever you are in your home. You don't need to pull out your phone and ask Jarvis to do something, he is always there. Jarvis is wherever you are, in the home, office and in your pocket if you use our Android mobile app.

Energy Efficiency

Every home has lights, appliances and other electronic or electrical gadgets. These all consume energy even if you aren't present in the room. Jarvis is smart enough to switch off lights and unused appliances when you walk out the room. If you turn on a light in your room, Jarvis will compare that specific lights consumption in real time and compare it to data from other lights. If it detects that the light you have on is less efficient, Jarvis will automatically switch that light off and turn on a more efficient light instead. How about automatically switching off cell phone chargers when your phone is disconnected? All these small things in human error or laziness add up in energy consumption and on your monthly bill. Jarvis is here to help make energy savings available to everyone without the owners even noticing. He logs data periodically and that data can be used to help reduce more electricity where possible each month. This way Jarvis will always strive to exponentially increase energy savings and decrease your bill.

Ease of Living

With Project Jarvis is your home, you no longer need to remember if your T.V or stove was left on. Simply login to the web app and switch them off manually or let Jarvis do it automatically. You can always check in on your home and turn appliances and lights on or off from anywhere in the world off any internet connected device. If you have kids, you can always check up on them. Are they doing their homework? No? Well simply turn off the T.V or their gaming console from your fingertips. Whether its by voice, gesture recognition, web app, computer app or mobile app, Jarvis has you covered.


Jarvis is capable of reading notifications, cellular messages, social network news feeds and so much more. He can even reply to those messages for you if you tell him what to say. Jarvis converts all...

Read more »

  • 8 × 5 Volt 10 Amp Relays Look for G5L Relays in EagleCad for pin layout
  • 1 × 74HC595 SMD Latch enabled shift register
  • 8 × EL357N-C Opto-Isolators. Any 4 pin, transistor output isolator will work
  • 8 × 2N2222 Transistors Discrete Semiconductors / Transistors, MOSFETs, FETs, IGBTs
  • 8 × 1N4007 Rectifier Diodes Rectifier Diodes
  • 20 × 560 Ohm Resistors +-5% Resistors
  • 20 × 220 Ohm Resistors +-5% Resistors
  • 20 × 660 Ohm Resistors +-5% Resistors
  • 8 × Blue LED's 5mm Blue LED's. Any will work but adjust your resistor value
  • 10 × Pin Headers 3mm Spaced Headers. Same as Arduino. Use what suits you

See all components

Project logs
  • Rough Artist Renders

    a year ago • 1 comment

    Check out our rough artist renders for our new Project Jarvis "Black Box". This box will have all the wireless transceivers used for the main home automation control and also have a lot more hardware and software for everyday lifestyle. We are incorporating sensors to track sleep patterns and monitor temperature, humidity, light etc in order to give you the best possible sleep every night. That along with a few more features will be added. 

    In the above image, the logo is a button. Obviously wont look like that but itl be tinted perspex with a soft glow through it. Thats the main power button when the device is off and serve other purposes when its on. The thing above that is a continuous rotatory switch so you can scroll through stuff on screen, adjust volume etc and if you click it in, it acts as the "confirm" button. Lastly there are a few buttons on top for any extra features we may add. Our designer's software got corrupt and wasn't able to finish it. A few changes to be noted is the lighting will be a lot softer and the red will be a soft white glow instead. The box also wont be so outstanding, our designer exaggerated the design a bit so the geometry wont protrude as much as in the design. You can check out the rest of the designs here

  • Build Details for Motorized Door Lock

    a year ago • 0 comments

    I've finalized my door lock and it's ready for the community to build. You will need a 3D printer unfortunately, however if you would like one, I can sell it to you for the price it took to make, a few dollars. 

    Anyway head on over to thingiverse and download the model. All instructions are posted there to help build the gadget. The model works very well and its very strong once you've locked your door with it. 

    This door lock was also featured on the official Makerbot website! You can read the short story here.

  • Wall Mounted Tablet Update

    a year ago • 1 comment

    Here's an update on our Jarvis wall panel tablet. I have mounted it on one of my bedroom walls and wired it for power however I still have an Ethernet cable to route for internet access on the tablet. I have a few more things left like building a power button and maybe putting some LED's. You can see how the tablet looks in the two images below. I have also made an application specifically for use with this tablet which can also be seen below.

    Here is the software I have built for it. I am going to incorporate Amazon shopping along with the grocery service they offer. Of course music for those early morning wake up tracks :D, but it will be a fully featured media centre that can stream from other devices. I'm going to do that later down the time line. I also have a space left for weather. From there you can see very detailed stats, a bit more information than what your iPhone provides and this will be able to display the info on screen, push to your phone(Maybe) and obviously tell you the info when you wake up via voice. The "Home Layout" tab will be replaced with "Home Essentials" and that's where weather and all the extra stuff can be found. I might also replace "Notes" with something more essential once I can find something I need. If I do, notes will be moved over to "Home Essentials". Notes basically stores all your notes, ideas, calendar events, etc. So far the home automation works, Music and Notes. Weather does also work however I didn't build a GUI for it yet. Next week I expect this software to be completely done and you'll be able to download the source. Our artist is also working on the home screen graphics which is why some are missing.

    NOTE: The app is built to run on the Beaglebone's LCD 7" Cape. The resolution is 800*480 and this app is designed to run in full screen. 

View all 28 project logs

Build instructions
  • 1

    The most important instruction, be very cautious! We are working with AC and you can easily shock yourself. Here are some steps to take when altering or modifying AC light / plug connections:

    1) Wear rubber soled, fully closed work or electricians boots.

    2) Switch off power to the room you are working in via the main distribution board.

    3) Use at least 1000v rated electrician screwdrivers. Do NOT use motor driven or other types of screwdrivers.

    4) Test the voltage in the socket using a well rated multimeter. Only proceed if you see 0v. If not, switch off the power from your D.B.

    5) If you remove wires, make sure you cover the uninsulated ends with electrical tape just to be extra safe.

    6) Once you have finished your modifications, make sure everything is well secured. You do not want wires popping out the next time you open the socket. Double check all your connections and then close up the socket.

    7) Switch the mains on and you can now test out your circuit.

    (Pictures coming soon)

  • 2

    The first thing to build is our relay board controller. These are simple to build because I've already created (Not yet tested) all CAD files which includes complete schematics and boards. You can find all downloads which are open source under the featured images of this project, look for the link to my Dropbox folder where everything can be downloaded and imported into EagleCAD.

    Pick your favourite relay controller! We have 4, some are cheap, some are safe, some are huge, some are tiny and one is a complete motherboard. The choice is yours but to help you choose the one for you, read below about more detailed specifications.

    Single Relay Module:

    This board is small and can fit inside a light/plug socket to be well hidden. There is no DC-AC isolation. This board is very cheap to produce due to its small square inch form factor. There is no shift register so you will need to make your own breakout board.

    8 Way Relay Board:

    This board is medium sized and has 8 relays on it. There is a shift register and you can make more boards and bridge them to get more outputs off the same 3 pins. There is no DC-AC isolation.

    Opto-Isolated 8 Way Relay Board:

    This board is slightly larger than the 8 Way Relay Board and has 8 relays on it. There is a shift register and you can make more boards and bridge them to get more outputs off the same 3 pins. There is DC-AC isolation. This board costs around $20 per board to make from OSH Park. This price only includes the board and no components.

    Jarvis Motherboard:

    This board is huge with 8 opto isolated relays and support for expansion. We have all the components and connections on this single board. This includes temp, humidity, current, voltage sensors and a whole bunch more. This is basically the complete package, less wires and its a nice single board. Coming Soon!

  • 3

    Once you have the board you like, send it off to OSH Park or your favourite PCB manufacturer. 

    While you wait, get the parts ready. You can find a parts list for every board inside an Excel file as well as a text file and that file only corresponds to that specific board. 

    The Excel document is well layed out but if you don't have Microsoft Office then just use the text file.

See all instructions


irinakim12 wrote 3 months ago point


I’m irinakim and I’m work in the Wiznet.
Thank you for using our product.
We are collecting so much data using our product.
And I will post the this project on our Web site.(
Can I post your product on our Web site?

Are you sure? yes | no

lionwang wrote 3 months ago point

banana pi support forum :

Are you sure? yes | no

AzizulAli wrote 3 months ago point

WOW!!.. Amazing work bro, your jarvis makes it look cool and brilliant. Hopefully i can work with you..:)

Are you sure? yes | no

abhijit.phondke wrote 4 months ago point

hello, how can i download the code for the app.

Are you sure? yes | no

cinderman wrote 5 months ago point

Dear IamTeknik,

I have been involved in home automation for several years now, and have developed a few components myself.

There are obivious aspects of this kind of system that I would NOT want to develop, when there are good, commercially  available designs/components already out there.

As far as "A.I." goes, I'm fairly certain that true AI does not yet exist. Most times you hear of this it is in a very "loose" sense. Usually the colloquial term is "smart".

AI is a very broad and complex field. I have not been involved in AI for several years but a quick search of MIT's CSAIL (CS and AI Lab) publications site brought up many interesting theses.

I don't mean to ramble, so I'll get to the point.

I've been developing software for my own "Jarvis" and the part that I haven't attended to is voice recognition. I have several ideas, just haven't commited to one yet. So naturally I explored your project hoping to find your solution to this component of home automation.

I cannot find any of your code.

This is supposed to be an open source community. If there are aspects that you feel should remain "proprietary" that's perfectly fine. You don't have to disclose ALL your work. But without more description or exemplary code, it's all, well, either a fantasy, or not meant for this community.

Each thesis related to AI (or any science) may take years to complete, and these are just one minor aspect of a broad field. I'm curious as to how you've managed to design an entire AI system in so little time.

And finally, I can understand not wanting to release "buggy" code, or "wanting to make the code easier to use" but that's not necessary. You release alpha code and everyone who is computer literate understands that it is possibly buggy and incomplete. They might even fix the bugs for you, and faster.

Again, I took the time to write this extensive comment because I have been involved in home automation for years and know there are still many obstacles to overcome before every house can be a "smart" house.

Please do not take this in a negative way but rather consider it constructive criticism, so your efforts can be useful to the community and proper credit can be given to you.


Are you sure? yes | no

Anand Tamariya wrote 3 months ago point

Voice recognition is easy with Google API. Have a look at my project

Are you sure? yes | no

cinderman wrote 3 months ago point

I realize that. But what if I don't want to rely on the Giant? What if internet is down?

Voice recognition should be an autonomous component, unless service is guaranteed.

Are you sure? yes | no

Anand Tamariya wrote 3 months ago point

There was a project Saera on N900 using Sphinx for offline recognition. You may try that out 

Are you sure? yes | no

Lucy Chen wrote 2 months ago point

From the looks of the page, it seems that the project is mostly a code for a HMI not an AI. There doesn't seem to be enough sensors and data collection inputs for a hypothetical AI to learn about its user. This home automation system is a Smart Home, basically provides control of multiple household functions and features in one attractive HMI. It is good for those who want remote access to their thermostat, lights, kitchen appliances, security system, etc.; however, this is already well in the market. AI like Amazon's echo and Apple's Siri tend to be short of inputs for more data acquisition points of the house. If you're interested in a touchscreen HMI...

This is a link to our touchpad controllers

It comes with HMIWorks, which is a free software developed for the touchpad. It is programmable with Ladder Logic or C Language.

This is the manual for HMI Works

I usually do demos at some shows with this technology. Follow me for updates on when and where I will do my next demo.

Lucy C.

310-517-9888 x120

Are you sure? yes | no

Kent wrote 2 months ago point

You appear to assume the explicit purpose of artificial intelligence is always learning about a human user. Reductionist view of the potential AI systems have in non-narcissistic useful applications, that. So we've decided Apple's glorified Eliza is the definitive implementation now? I get that there is convergence between what Apple's shooting for and this ... but OTOH, Apple just pulled together stuff people have been working on since the X-10 days and put it in an app. Who's to judge?

I'm not opining on this project in specific, but it's notable that the active global-scale push to model a human brain began with creating a system to replicate a single neuron. Sure, the brain has millions and millions of neurons - so just one neuron couldn't possibly be a brain. That doesn't change the nature or objectives of the project.

I'm very new here, but damn this post is spammy. Is this how it goes at Hackday .... people show up on other projects essentially to bag on them and hawk their own commercial offerings? This feels weird. 

I thought we were supposed to be helping each other make our OWN touchpad controllers? Obviously, no matter what project an individual hobbyist decides to attempt, a funded corporation would almost certainly be able to do it better. If that's the only point, at least for me, than there's really no point to any of this at all.

Are you sure? yes | no

Todd wrote 6 months ago point

I'm new to site and tech ignorant so please forgive.  Why can't the xbox system with voice commands be hacked and improved to be implemented into a home.  Feature allows to turn on and off tv/xbox as well as volume control and internet (speech recognition) searches. 

Are you sure? yes | no

Anand Tamariya wrote 3 months ago point

A hacker can hack it but can't publish the source code as it's proprietary. Then it won't be reusable by others.

Are you sure? yes | no

sohaib_1090 wrote 6 months ago point

this project should be more concentrated on software side rather then hardware.and the author of this project wasted his time and putting his many months effort on building a inexpensive easily available relay module and a simple door lock. and he is now again in 2015 hackaday prize........,i think this time the author will come up with a microphone module. and  2015 progress will be  a microphone module schematics and parts list.i hope the project will be completed in next century and will win 2115 hackaday prize :-p

Are you sure? yes | no

IamTeknik wrote 5 months ago point

I apologize for the delays but i am a university student and I do run my own local business so it's difficult to find time to work on my own personal projects. Also regarding the existing technology, you are correct however I do plan on making kits available or maybe making a commercial product, so it is essential to have these modules made.

Are you sure? yes | no

flare wrote 6 months ago point

I thought this was abandoned. Is it a new project Jarvis, or the old one being picked back up?

Are you sure? yes | no

-= Wolf =- wrote 7 months ago point

dude, this is awesome, i've been planning on building my own system to automate my mancave and eventually the whole house, i flearned to make android apps and have it interact over bluetooth with an arduino, and some other stuff, but i just haven't had the time to do it and work on it, you inspired me to pick it up and get it to work, good job and keep it up!

Are you sure? yes | no

Arunan wrote 8 months ago point

I'm so impressed. Am very much interested in your project. I wanna get the same.. 

Are you sure? yes | no

Serious Nonsense wrote 9 months ago point

I interest your project.
Give me your code please!!

My E-mail is .

thanks you!

Are you sure? yes | no

Quad Brain wrote 10 months ago point
Great work!
When can you share code files? I wanna make same.

Are you sure? yes | no

shan@tek wrote a year ago point
Hey it very nice project but when u r going to upload whole project with codes??? I m from india I want to do same in my home

Are you sure? yes | no

seesemichaelj wrote a year ago point
I'm confused why you're building all the devices for the automation? This has nothing to do with the actual home automation. There are several door lockers, tablets, tablet interfaces, etc. The only hardware you should be building should be for speech recognition, energy saving, and AI home automation features specifically. Frankly, I think the only thing you should be developing is any hardware that is not available COTS (Commercial Off The Shelf) and the AI component of the software. Home automation has been done and done again. There are several open source projects too (i.e. OpenHAB/SmartHome) that you could incorporate with the world. As a hacker, I think you're wasting a lot of your time on things that should not be high priority. This project is supposed to be open, so why haven't we seen proof of concept design on how the AI is going to work (the only original component in this project). As a business-man, your time-to-market is very slow. You should be making partnerships with other manufactures and software companies to integrate with their "smart home" systems. Instead, you're going to make an entire home automation system that has it's own configuration, devices, features, and bugs. I'm sorry to be critical, but I really see no progress as intelligent system (your videos demonstrate no AI); I only see progress towards a home automation solution.

Are you sure? yes | no

seesemichaelj wrote a year ago point
I guess there's no edit feature. I meant to say "This has nothing to do with the actual project you set out to do, which was to incorporate AI into home automation." If your original goal was just home automation, then you should not have submitted to THP.

Are you sure? yes | no

VPugliese323 wrote a year ago point
While I did not go through your entire project page, I was not able to find any documentation on the theory behind the individual sub-sections of your project. If it does not exist, I suggest adding it. If it does exist, then I suggest you make it more prominent. Perhaps its just the computer scientist in me, as the first thing I look for in a project is project objectives followed by project theory in an abstract form. You have excellent descriptions of what the different modules do, I just don't see any of the theory behind them clearly laid out.

On a different note, I could not find any means of diagnosing faults as a part of this project. I suggest you add a diagnostics engine (of some sort) to the project. Having any kind of diagnostics engine (even a rudimentary one) will aid you significantly in error handling. Something is bound to break. If you do not want to code one up from scratch yourself, I believe there is one that you can modify from the robotic operating system (ROS) found here:

On a final note, as I scan through your project logs, I see a lot of good progress. This looks like a nifty project. I look forward to seeing the end result.

Are you sure? yes | no

Big Country wrote a year ago point
Are you ever going to release your code? It looks like you are trying to enter a closed source project into an open source competition.

The competition is supposed to spread ideas and spark an open forum to discuss and open up tools that people may not have been inclined to share before. Thus, flooding the hacker community with helpful information on tools that people have been developing behind the scenes. So far, the things you have shared can be readily found on HAD prior to your posts.

Your initiative part of your project is without a doubt the code that makes Jarvis tick, if your claims are true, it rivals many other commercial products created by large companies. However, nobody can substantiate your claims because nobody has seen your code or even ran your program. Your website still looks like you are trying to sell this program as a commercial product, which contradicts your replies below that you will release the code "soon" over the past 3 months. I don't think the readers will be distracted by your reiterations of other projects to hide the fact that you are hiding your code. If you don't want to release your code, that's OK, but please don't market it as open source or string people along with false promises.

Please prove me wrong, I sincerely hope that I have the wrong impression of this project.

Are you sure? yes | no

IamTeknik wrote a year ago point
Ok to address the smaller things first. Yes this will be a commercial product eventually but what is wrong with that? Arduino is a product but still open source, it is exactly the same thing.

Secondly, a lot of my builds are relatively new. Yes the basic automation is pretty standard but i have a lot of other things in the works. The door lock, fingerprint scanner, wall mounted tablet etc. thats just what ive done. I have a bunch more stuff which have never been done before. All of which will be posted once my designs are finalized and working well.

My project is marked as an incomplete project which is very evident. Im one person working on this which is by far the biggest project compared to anything else in the semifinalist list and I'm still a university student so i apologize if it takes longer than you would hope to have a fully functional bugless A.I. Its a lot of work for one person in the time span I have and unfortunately I do have other priorities which I cannot neglect.

The code is not yet released because it is simply not at the level i want it to be however i will be releasing some code next week which will not be the A.I but rather a substitute that people can use in the mean time. I will not release bugless, broken code for a community to use. There are companies with billions of dollars, 1000 man teams and quantum computers trying to write a great A.I so 3 months to write a decent A.I is a really short amount of time. It may not be the overnight solution you want but thats how things work, it takes time to develop something new.

I think you are out of place to say its a "fact" I'm hiding code just because you haven't seen anything in 3 months. I don't build reiterations of anything to hide anything. I build because I want to build and i will continue to do so.

Are you sure? yes | no

zakqwy wrote a year ago point
I have to concur with Big Country's point. Please don't take it personally, but we need to see your code. No one will fault you for showing off a program that isn't polished; just include a few good caveats in the comments (many open source licenses include an 'As-Is clause either way). I think the point being made here is that while your electrical parts are useful, they aren't revolutionary; Arduino-controlled relay modules, fingerprint interfaces, tablet wall-mounts, and various rapid prototyped enclosures are great, but they aren't that different than existing hardware.

The real magic of Jarvis was shown at the beginning of your Hackaday Project Phase 1 video--your AI clearly interpreted a statement based on an understanding of your intention, not just a translation of your vocabulary. Even in its state when you produced that video, the code that produces that functionality really could be revolutionary. That is what we are all dying to see. Show us how Jarvis thinks.

If you're committed to making this project truly open source, I challenge you to properly license your code and release it in its current form, with any caveats you feel are necessary to communicate the fact that it is currently a work in progress. You should be able to do this in a few minutes--call out the specific license per its requirements in your comments and copy everything to a Github or Dropbox repo, then share the link on the project site.

Are you sure? yes | no

leocoins93 wrote a year ago point
I need help on how to get started PLEASEEEEEEE HELP ME . I have a Mac is it compatible ? where is the software? Please help me.

My email is I am very interested in knowing about this project.

Are you sure? yes | no

Bruce Cushman wrote a year ago point
I love your concept and will be following your work closely. I am currently retrofitting my 1909 Craftsman home into a "Jetson, home of the future", and would love having a Jarvis system to control all of it's features. One question I have is, why didn't you just use Arduino relay boards that are available, very cheaply, on eBay. They typically cost about $1 per relay and come in 1, 2, 4, 8 & 16 relay boards with opto isolation? Also, an electric door lock ( 12v, made of metal) can be had for about $25 that installs into a standard door jamb.... although I do really like the look of your rack & pinion lock. I don't have access to a 3D printer at this time so I need to be able to buy "cheap" off the shelf parts for my Jarvis home.

Are you sure? yes | no

dklight wrote a year ago point
Grate work! A while ago i tried to build some home automation system. A very interesting side is the feedback. Take a look on since it's an open source project, and everything is very well documented.

Are you sure? yes | no

zakqwy wrote a year ago point
Can you put up a link to your code (or point me in the direction of the firmware if I missed it)? I'd love to get a better understanding of the software that backs up the functionality demonstrated in your 2-minute video. I don't know much about voice recognition but it's definitely cool!

Are you sure? yes | no

IamTeknik wrote a year ago point
I am going to upload the code soon, im just working out a few bugs and adding more functionality so someone can get using it without too much problem

Are you sure? yes | no

zakqwy wrote a year ago point
Any update on sharing the code demoed in the video?

Are you sure? yes | no

Similar projects