Show HN: The home automation system with an API
new.ninjablocks.com
new.ninjablocks.com
2) Like Twine, another HA startup, and all of the incumbents, it's too expensive.
Then you have a startup like SmartThings (http://smartthings.com/) that's just reinventing the wheel, creating a pretty iOS interface, and then trying to charge an outrageous amount for a monthly service - something MiCasaVerde doesn't do. So, at the end of the day would I use SmartThings? I'd like to - but I won't because I don't want to pay $120 a year for something I can run better in my own infrastructure. I don't want you selling my energy usage data or knowing if I'm home, with something like MiCasaVerde - I can choose not to use their cloud proxy to control my home.
If anyone is listening who's designing home automation and wants to cater to people who are looking at building a service to deploy these types of automation to the masses you need to do it right. Baby Boomers, at least the majority of them, will think all of the above are too complicated, so they're going to want someone to install and maintain this for them. There's a huge middleground that needs to be solved here and unless SmartThings and others are going to provide a service, that $10/month is not going to be justified.
Back to NinjaBlocks, please answer the following: 1) Monthly fee if I use your cloud proxy to hit the controller? If so - how much? 2) If #1 is true - is the API directly accessible on the controller - or are you locking me in? 3) Low power integration (i.e. ZigBee, Z-Wave, Insteon, etc.)? Answered my own Q: No... From (http://help.ninjablocks.com/customer/portal/articles/692139-...): "The original Ninja Block comes with a RF 433.92 Mhz (433) dongle and future Ninja Blocks will have this baked in. 433 is awesome because you will find it in all sorts of low-cost battery powered devices, its range is pretty good too. That said, its not all sunshine and lollipops - there are good reasons its considered "low-tech" - only one device can "talk" at a time, labelling is appalling, all protocols are proprietary. So not everything will work out of the box, but a lot of stuff will. " <<<<<----- Sorry, but this sucks, I'm out. You need to have a real and baked solution besides this 433Mhz junk that's not mesh, and has zero security baked in. There's no way I'll pay $200 for this with zero long term scalability or security in mind.
Again, another pretty solution that's just a toy. For my money I'd go with SmartThings - but for the long term I'm sticking with my MiCasaVerde (http://micasaverde.com) setup - cost feasible and fully in my control and no monthly costs. Until someone can come to the table with that I'm better off building my own interface.
And, Z-Wave isn't emerging - it's been around for quite a while. It's not perfect (most low power mesh wireless networking standards have their flaws), but what NinjaBlocks has chosen to implement is not something you want in a system you depend on. Insteon has the advantage of wireless plus being able to operate over powerline (i.e. your home power distribution).
It's a bummer - I've gotten my hopes up for a few platforms as of late, all have been let-downs in their own way.
Ninja is ultimately (read: not yet) a meta-platform for connecting stuff. As you point out, there are plenty of different approaches and technology with advantages/disadvantages. We think the thing what is really missing in the landscape today is interoperability. Say you want to give access to your security system's motion detectors to a third-party app that helps you save energy, that is nigh on impossible with an OTS offering. That kind of thing is exactly what we enable (albeit a little crippled) today.
Its early days and we have a metric-shiton of stuff to do, some of which includes:
* making the webservice fully or partially optional - current cloud driven architecture is for expedience * supporting lots more sidebands than just 433 - again chosen for expedience and cost. Z-wave is better, no doubt, but its not really justified in the mega premium per device i.e. $70 for z-wave PIR vs ~$15 for 433. BTW we have proof-of-concept support for Zigbee Light Link via a USB dongle on a block. * publishing a public repository of local modules to make it easy to wrap (and share wrappers) for proprietary/lan based offerings (think wemo, philips hue, etc)
Another thing to note is that we are a hardware company by necessity not choice. You can run our (node.js) client on an RPi, beagle bone, media centre pc, etc.
Anyhoo, I digressed. We hope to have something that you can get excited about one day in the not too distant future ;)
I think the NinjaBlocks folks are missing the point here...
Interop is key in an automation / security / energy platform. Saving energy, as you state, is fully baked into the MiCasaVerde platform (and others I'm sure) today - see ERGY (http://www.ergyenergy.com/). So, please don't make that assumption that it's not there in others.
I still don't get your target market. It seems to me it's the people who don't really mind a half-baked solution, or is it geared towards people that will go the remaining 80% (there are a lot of things missing the more I look at this platform for a serious contender in full HA). Your cost estimates are way off base here - I've never paid $70 for a PIR with Z-Wave - try more along the lines of $35 and the cost of these controllers is coming down. And, again, you can't take 433MHz seriously. You do realize that your entire system is proximal to the RF of your 1 controller, right? Whereas in my Z-Wave system I can fully extend out into a huge building and take full advantage of a self-healing, self-routing mesh network. As I stated before - 433MHz stuff is junk, and if you're even paying $10 for a similar PIR controller, you're getting what you pay for - something that doesn't work well. This is why I'm not a fan of WiFi only devices such as SPARK. Think very large home or business - unless you have great WiFi coverage everywhere (since these little devices generally have crappy radios / antenna) . This is something people who don't actually use home automation won't think about - because they don't actually use it. But in my mind 433MHz in an HA system is akin to Internet banking without encryption - sure, you can do it, but it's going to break quick once you really want to take it seriously.
So with the system I have implemented today: * Control Insteon * Control Z-Wave * Control IP devices * Local API access
The things that are missing are: * Control Zigbee (this is the biggest miss and I applaud SmartThings for picking this up) * Better UAC
This isn't a dig, but instead of making a lot of the assumptions you've put forth in your response, you should potentially team up with people who are seriously using it today and look at how they're extending it in their homes. You'd learn a lot I feel, and actually put out a product that addresses the same pitfalls everyone who's trying to play in this space is making. I know a lot of people who are using this stuff today - and we all have the same sentiment for all the new "players" - but at the end of the day the products are all still not serious contenders and I'm baffled by how SmartThings pulled $1M+ from Kickstarter by reinventing a prettier wheel.
Good luck on your attempt here - and you may want to look at what's out there in the mean time.
Would love to chat offline if you were interested in hearing our take and/or brain dumping on us - pete at ninjablocks com
(Ninja Marcus here) Thanks for your feedback and questions!
1) Nope, no monthly fee. 2) We don't support this today. We think most of the value actually occurs in the cloud. Having said that all our APIs, hardware and code are open source, if you want to implement this you can. 3) The question of protocols (433Mhz or otherwise seems to miss the point of Ninja Blocks (& the Ninja Platform). We want to integrate with ALL connected hardware. We've been working on adding other protocols like ZigBee, Z-Wave, Insteon but they're not ready yet. I personally believe the future is IP.
We've included 433Mhz because it gets people started very cheaply!
Cheers,
Marcus
Thanks for your response. I'd, kindly, disagree with you on #2. I'm a serious home automation user - and have it deployed in multiple locations for myself and family. I won't buy a controller that banks it's existence on Internet connectivity - there are use cases for having controllers air-gapped, and in my book forcing an API in the cloud is a big miss. Hope you'll reconsider, until then - you won't find me as a customer.
'value' for you occurs in the cloud, to me it is a privacy/security risk waiting to materialize in some form and it feels grafted on to create a subscription model where a one-time payment would suffice.
Not sure if people who use PayPay for credit card processing are aware of this. I'd love to buy this - and I attempted to - but can't.
(I'd love to use Pin, but I'm still waiting for a way to use them without the massive bugbear of anything beyond PCI SAQ A, which comes into play as soon as you're POSTing CC numbers to your server.)
PS: Hi Marcus. ProgSoc Rob here. :D
(BrainTree just launched in Australia[3], though. I'll be checking them out shortly for a project of my own. :) )
1. http://blog.angrymonkeys.com.au/why-being-an-aussie-startup-...
edit- It does not support the sensors :(
Also, will it work with standard security/video cameras (RTSP/RTMP)?
edit- no, it does not look like that works now either :/
IP cams (especially the ultra cheap ones with pan and tilt) are way up on our list. It will be one of the first client modules off the shelf in the new year.
http://micasaverde.com/vera-lite.php
The Vera Lite is probably what you want... It can handle your existing sensors and do video (although the video isn't anything spectacular - I use a different system for that, but it does work).
The Vera has "app" concept - so from the controller you can install MiCasaVerde or community provided applications to do certain things (i.e something like a countdown timer for a scene, or an app to control an OTS yard watering system for example). But the best part is you can write your own apps (LUA). The system is entirely hackable (SSH access). So it's basically an open slate to most extent. But, the key thing is there's no monthly fee if you want to use their cloud service - you just need an account and controller (I'm guessing they collect stats on you events - but that's just a guess on my part).
The UI leaves something to be desired - and again, that's where startups like SmartThings will capitalize (pretty).
I am as convinced as ever that almost all home automation needs, realistically, are covered by a WiFi-enabled drop-in replacements for conventional wall switches, junction boxes and sockets. Everything else is nice, but you must get the basics right before moving on to the fluff. And nobody still has got it. I want to be able to switch off a floor lamp 5 feet away and turn up the heat in the bedroom without lifting my precious behind from the couch. Whether there's a motion in a garage is really secondary.
We do remote control sockets today. Not elegant but at $10-15 a piece best priced laziness available.
We are cooking up something for IR controlled heating/cooling.
The IR controlled heating/cooling ... IR? How is that supposed to work?
In Australia, central heating/cooling is rare and high-end - in most places cooling matters much more than heating. The most common climate control method by far are split system airconditioners. We are working on a retrofit device specifically for these, though it will work on any IR controlled appliance.
Wifi is subject to a single transceiver - that does not automatically extend. This is a big problem in tight spaces, because generally the smaller you go, the worse reception / transmission power you'll have in these systems. What you want is a mesh network (i.e. ZigBee, Z-Wave, Insteon, etc) that will automatically handle the mesh aspect of being able to continually extend itself as you build out. It is also much less maintenance long term as well (for a variety of reasons).
But if you think that there aren't drop in replacement wall switches already - there are. Aeon Labs (http://aeotec.com/z-wave-in-wall-switches) makes some killer stuff that is in-wall, and there are a multitude of others. Check it out - everything you want is out there. And you can build an entire HA system in a decent size living space for far less than $750 USD.
That's a special case of a more general solution. The best way to solve a problem is to not have to solve it.
Living small also reduces the need for fancy power usage monitoring, WiFi repeaters, etc.
Your homepage says the Ninja Block contains an Arduino-compatible µC and a BeagleBone. It took me a while to understand that the former is for communicating with sensors and the latter is for communicating with your API. I didn't realize they worked together at first - I thought it was some sort of "we used Arduino initially but then had to upgrade to something else in later versions" thing.
Interesting use of USB connectors to pass power and IO to accessory boards.
Thanks for the feedback, we should make that more clear. Actually both the Arduino-compatible µC and a BeagleBone will be used to talk to sensors.
Having a tiny linux computer on board buys us a heap of flexibility!
Cheers,
Marcus
We live in an on older (70ies) apartment building, with a wonky central heater, and people with varying levels of tolerance for heat and cold. Some of us suspect the heater is kind of broken, because it's hot when it's warmer out, and sometimes seems to pump out less heat when it's cold.
So... it'd be interesting to collect data: interior temperature for all 6 apartments, and exterior temperature in the shade. And do so cheaply. Ideas?
pros - You can retrieve folded historical data for sensors from the api (min/max/avg/interval etc) - You can set a callback that POSTs every reading (every 45s) - cheap, additional temp/humidity sensors are $15
cons - range could be a real problem. Manufacturer states 100m open air 30 inside, which is inline with our experience (with a fully extended antena). Apartment walls could be a showstopper.
I'm sketching out the design for a raspberry pi-based central heating controller with a web interface, mix of sensors and an lcd display - should cost around 75-85UK pounds, including the pi
I work in HVAC controls and I'm always surprised by how the 'home automation' folks seem to just ignore what's already out there.
My favorite standard is BACnet (standard for ASHRAE, ANSI and ISO), but there's also Modbus, Lontalk...
There's a bunch of people who have already worked very hard to make sure there's some usable protocols for lighting, air conditioning and even security access.
IMO you could leverage a strong position and use existing codes by using them.
The Ninja is a meta-platform that offers a normalized API for disparate hardware. See my reply above for the long winded answer. The short version is - we aim to provide OAuth for your all stuff ;)
The last control company I worked for was heavily invested in the Delta controls' products. (All working on top of the BACnet protocol) Last I heard Delta also started to toy with Zigbee.
In my head there aren't mutually exclusive, as the BACnet architecture enables it to be built on top of other protocols. (BACnet/arcnet, BACnet/MSTP, BACnet/Ethernet, BACnet/IP...)
You see, home automation seems to always do "it's own thing" in regards to industry standards, which is why I manifested myself here.
If you want to know more about building control however, you can just type "hvac control" in google and you should be in the right spot. https://en.wikipedia.org/wiki/Building_Automation
Cheers,
Marcus (Ninja Blocks)
It was 5 a few hours ago, then 12, now 19.
Pretty sure it is supposed to go the other way.
It uses KNX / EIB Bus industrial standards for Home automation, as well as enocean and other 3rd party automation tools.
And it's fully customizable / programmable with php.
Why isn't the "remaining" number going down?
A. Slow web hooks. The page is written in Sinatra. The sinatra app stores an inventory value in MongoDB.
When there is a new order, our commerce engine Shopify (is meant to) send a web hook POST to the sinatra app to get it to pull the current inventory via Shopify's API. This hasn't been happening!
Cheers,
Marcus
I'd love to talk to you guys about what your doing and how maybe I can integrate some high end gear with it.
With that, what are some some features users would like to see?
(Ninja) Marcus
https://s3.amazonaws.com/ninja-assets/gears.svg
:)
On the other hand, your copyright year is already in 2013.
Cheers,
Marcus (Ninja Blocks)
"Shipping Mid January 2013"
"First batch ships December 2012"
Cheers,
Marcus
-- Production delays -- I waited almost a year longer than expected to get my unit. Because I just wanted to tinker I wasn't bothered. Ninja Blocks say their units are manufactured, just waiting for the enclosures. This might not be a problem for them.
-- Battery eaters -- Twine eats batteries like nobodies business. Two AA batteries last about two weeks. On the forums this seems like the typical experience. Ninja Blocks looks as if the central 'Ninja Block' can be mains powered. But will the wireless sensors have a short battery life?
-- Web app -- Twine has a website that the device must communicate with to report data and push rules to it. There's no (out-of-the-box) way to get it to talk to a local server via a clean API.
This was a big disappointment for me mainly because their web app is very limited: - No data logging (just reports the latest figure) - Rules can just be triggered over thresholds (ie when temperature reaches 20c then X) not when temperature changes - For moisture the actual resistance isn't reported, just a wet or dry condition. This renders it useless for plant watering.
Twine can make a HTTP request when a rule is tripped. I plan to play with this but it's not that useful if it can't just HTTP request every time a value changes as I wanted to log temperature (plus it's a GET!).
The Ninja Blocks apps look better and appear to fill the gap between reading something then doing something. Twine can really only report something (twitter/sms/email/HTTP supported). But I'd caution that until you get your hands a device like this you don't know of its true limitations.
-- Sensors -- I didn't up my kickstarter amount when they announced new sensors. I wish I had, as only the temperature sensor is nearly useful. The other built in sensor is orientation... ? IMHO they should've thrown in the other sensors and maybe upped the price to increase satisfaction.
I prefer the sensor pack that ships by default with Ninja Blocks.
In spite of the above I'm optimistic about Twine as it really is early days. This space has massive potential if it can break out of communities like ours into the mainstream. I just hope Supermechanical doesn't get distracted and focusses squarely on their flagship product.
Ninja Blocks? I won't be ordering but will keep a close eye on what those who receive the first units think. Their approach so far seems to be a better path to market than Twine's crowdfunding and their presentation superior.
-- Production delays -- this is our second go around, we had a kickstarter back in March and it was late (and not a great product to be honest). We learnt a lot this time around and ponied up for a small run. We were "sure" we would ship in December when started taking orders on the 11th (of Dec). Alas, Murphy hates us. Final assembly http://www.youtube.com/watch?v=ZQKQw-YXyjA is happening now. Bleeding into the holidays has cost us a couple of weeks, 15th Jan is about as iron clad as anything can ever be thats outsourced to Shenzhen.
-- Battery eaters -- the block (a beagle bone + arduino cape) ships with a 3A switching power supply - no messing around. The 433 sensors are off the shelf items you find in low end security systems. They use most of their power when they transmit (which isn't that often) and (whilst not tested) we estimate the batteries will last a long time.
-- Web app -- We have a web app that lets you see live data and make simple rules. As you have probably figured out with Twine. Whilst that is fun and sometimes useful but pretty limiting. We think making your devices available to third party apps via OAuth is going to make things much more interesting!
We have mountains left to do, but we already have:
- inbound and outbound webhooks, which enable simple integrations with stuff like Zapier or Tasker no code affairs. - a fully featured API that we use for everything, i.e. not an afterthought- see http://docs.ninja.is - separate interfaces for creating user land apps and devices - OAuth2 - Open source client (currently being refactored to support third-party modules a la npm/gems/etc) - see the "release-two" branch https://github.com/ninjablocks/client
-- Sensors - lol re-orientation, we had an accelerometer in our kickstarter block which was removed precisely because no one used it. Pretty much the first time I found an real world use was for our friday afternoon rube goldberg machine http://youtu.be/jPE_M0ciCYI (beware shaky cam, but we weren't doing another take)
We really wanted to support remote controlling power sockets as soon as possible, but at a reasonable price. That eventually led us down the path of 433 as the only globally available option. Insteon, Z-wave, even Zigbee are "better" technically - but they are 50-100 per socket vs 10-20. We've secured a CE certified supplier for Europe and Australia and will be reselling soon for $12 a piece. There are ok options on Amazon for the US and we are lining up something for the UK - we can't justify the MOQ just yet.
We opted for off the shelf sensors the second time around for a whole bunch of reasons (see above). Cost, availability, but mostly because hardware is hard and oh so slow - especially for software guys. Note that we think the value is in wrapping everything, especially all these awesome kickstarter/indiegogo wifi devices that will probably have average APIs (no offense, but hardware guys just don't seem to grok APIs or auth - think wemo ... LAN only undocumented SOAP, really?)
There are important things we need to tick off - the huge one being severing the link to the mother ship. Currently all logic lives in AWS, thats not acceptable long term. Unlike some of the people in this space, we want our users to connect to our service because it adds awesome value through an ecosystem, not because you dropped a bundle on their hardware which requires it.
Important note - you can already make your own devices. We've got an Arduino Ethernet library that runs on a 328p (that's 2k of memory for the uninitiated) https://github.com/ninjablocks/arduino-ninja-blocks. There is even a browser implementation that uses cors https://github.com/ninjablocks/browser-ninja-blocks. Of courese you can run our official client (git link above) on anything that supports Node.js or just use the REST interface to make anything with http a device.
Sorry about the verbal diarrhoea - midnight Saturday in Sydney :/. I certainly don't want to suggest that the Ninja platform is all sunshine and lollipops today. However, if you a software hacker and want a real API for things. That's where we are going and we'd love for you to come along for the ride ;)
Again - way off base here - please do your homework. A plug-in on-off socket can be had for under $30 (http://www.ebay.com/itm/Wayne-Dalton-HA-02WD-Wireless-Small-...). The cheap 433MHz argument goes out the window very quickly if you know what you're doing.
Even the more expensive models are half the price of the Zigbee devices!
I've said it multiple times in this thread - you get what you pay for. The 433MHz (not Mhz) gear is non-mesh. If you don't currently have home automation I wouldn't expect someone to realize how much of a key this is to a reliable and useful system. But, go ahead - try running more than a dozen 433MHz devices on a controller - it's going to be unreliable and have restrictions based on proximity to the controller which will depend on the NinjaBlocks RF design as well of each of the individual plugs RF design. For $3, I bet it's stellar!
#dontsayididnttellyou
Our hardware engineer did get it working on Telstra 3G using a USB 3G adapter. It does require you to jump into linux land to set this up.
Cheers,
Marcus
Cheers,
Marcus