Hacker News new | past | comments | ask | show | jobs | submit login
Introducing The New IFTTT (ifttt.com)
263 points by johns on June 20, 2012 | hide | past | favorite | 85 comments



Note to hardware hackers: IFTTT almost certainly started with some musings about "wouldn't it be cool if I could make my household X trigger web API Y" and for many entrepreneurs would have led to a startup about smart gadgets. There have been many over the years.

The very clever thing about IFTTT is they sat and cheaply began the process of building the web layer.

I like it!

WeMo is likely the first of many many gadgets and interfaces where they can accept links.

Now they wait and see how quickly the China Ecosystem and the gadget makers like Belkin deliver lots of nice cheap easy trigger devices.

BugLabs...Nest...Dropcam...LG Smart Refrigerators...Toyota Prius...


It would be more exciting if they had an API.

I don't see any way to add channels - either a hardware one or a software one. You could use RSS but that is it. And I don't see any way to add an action. They could very easily add a "web hook" URL for a custom action.

I get really annoyed when I see sites like this that anoint winners. They already have an API to work with outside sites but they just don't publish it


I started last week to work on the API. Stay tuned.


Yes! If IFTTT offers an API for channels and triggers, then a whole world of enterprise software opportunities opens up. Sign me up!


Exciting stuff! That seems like an interesting way to bring back in the hardware possibilities from the comment above - using an ifttt API in tandem with smart devices could eliminate a ton of work in developing automation.

Also, what's ifttt like to work for so far?


Very fun and small team of 8 with great vision. Starting the week before this redesign was a little hectic, but it was worth the effort. Only big things to come :)


+1 on that.

If you open it up so that j. random startup can announce "Hey, we provide trigger foo, trigger bar and action baz" in a way that doesn't require anything more than cursory human interaction on the IFTTT end of the transaction, you suddenly become the default aggregation point for new services...

And may I say that having to worry about spammy APIs is kind of an awesome problem to have.


> I get really annoyed when I see sites like this that anoint winners. They already have an API to work with outside sites but they just don't publish it

It is much more important to get the experience right with one or two small specific cases first - and then to make it available to the larger public.


Yeah- a lot of startups are born from people coming up with an idea to solve a common problem... or even just an idea for something that they think would be neat/amazing.

Plus, they essentially are a user by default of a product/service they wanted for themselves in the first place. It's a great way to bring passion to the masses...


This is a little off-topic but I would still like someone from ifttt to answer in more detail the method that they use to store passwords for 3rd party sites.

I think this is an important detail for an app like ifttt as they necessarily have to store the actual passwords in a form that they can reverse encryption (for websites that do not support something like OAuth) for in order to submit on the user's behalf. So it's important to get some transparency on the mechanism by which ifttt does this.

I also posed the question here a while ago: http://www.quora.com/How-does-ifttt-securely-store-passwords...


> I also posed the question here a while ago: http://www.quora.com/How-does-ifttt-securely-store-passwords...

And you appear to have received an answer there, as well. Anything using OAuth doesn't need to store passwords, and for anything else ifttt stores the password in an encrypted (but obviously retrievable) form. I don't see how they could do otherwise; at most, they could refuse to interoperate with any sites not supporting OAuth, and I don't see the benefit to that.


To be clear, I was looking for more detail around how they were doing the reversible encryption.


People who know seem excited about this, so I spent two minutes following links. I still have no idea what ifttt is.

This would help:

- Change the tagline from "Put the Internet to work for you" to something more descriptive.

- Have a prominent "About" link on the blog. Write an About page, or better put some content on your home page.


I'll work on getting better info on the blog. In the mean time, here's our about page http://ifttt.com/wtf


When I saw "If this then that" I instantly knew what IFTTT was. However, when showing it to people I noticed they "got it" after seeing the recipe "If the forecast calls for rain tomorrow, send me an SMS message."

BTW, really really cool!


Even that needs a lot of work.

Okay, so I can create "connections" on an "if this then that" basis. I need more examples of the sorts of thises and thats that I can use as input and output, though. Like, apparently I can control my home appliances? And... what else?


Really? The second heading on the page clearly gives five examples with an obvious link to 46 more.


Why do I need another website to do Twitter, Facebook, Evernote, Email, or Weather? I already have apps that do whatever it is those things are.

No, the page in question does have 1 really good example, the if Instagram then Dropbox example. I think the homepage of IFTT would benefit from a nice animation where the "this" and "then" options cycle through various options. It would show how any tigger can be combined with any action and highlight a few of the good use cases.


Try the "Browse" link at the top to see recipes other people have made. http://ifttt.com/recipes


And what's with all the MASSIVE FONT?


SRSLY.

I have a highly hacked local user CSS stylesheet to get around the problem of small fonts on too many sites. I rarely have to resize pages, but this one was actually too big.

Something of a novel problem.


I don't think any of that would help. Right now, either you get the concept or you don't.

Further down the line when "normal" people will be adopting this (they will) then the recipes will be the primary selling point IMHO.

Just tak a look at a few of those and you should "get" what it's about.


I don't think the problem is he doesn't "get" the concept.

The problem is that the site doesn't even say what the concept is.


I have been using it from the beginning where the concept was even more vague.

I think you either get it or you don't. Apply whatever abstraction of "get" as you want.

There isn't that much to get really. If you are a programmer if this then that should kind of tell it all.

I am not a programmer (but know how to program a little) and I had no problem "getting" it. I don't believe developers dont get this.


I don't think you're "getting" their argument. The point isn't that they are incapable of understanding IFTTT. The point is that there is an entire blog post about integrating IFTTT with devices, with no description of what IFTTT is.

Reading the brief description on the About page or playing with the site for 2 minutes would solve this, but neither of those things are discoverable from the blog post.


I think I am getting their argument and I find it to be missing the point.

A blog in this context seems more to be used to tell people about where in the process a given company, service, product is.

If you are reading the blog there is a big chance that you "get" the product already.

Anyway. It's not my business. And I agree they could have an about up. But it seems to be a problem way down on the priority list. That's just my opinion of course.


There are many ways to get to a blog page - Google for one, Hacker News for another. You really need some way to put it into context. Why lose the opportunity to introduce another potential user to your service?


So what you're saying, is every blog post written on a products site, needs to re-introduce the product each time? That's a fair request actually, and could be achieved as simply as an opening sentence of 'If you're new here, you might want to learn about us first! CLICK HERE'


No, I'm not saying there should be explicit text in the blog. I'm saying there should be a link somewhere on the page that takes you to the home page or "About Us" or "More about <product>". That's good design practice for any web site, because there's no telling how someone landed on your page. Finding a leaf node shouldn't leave you stranded.


Well, there is a link back to ifttt in the very first sentence of the blog, but something in the header or sidebar couldn't hurt I suppose.


IFTTT will never catch on with "normal" people because as cool as the tech might be it's a Internet-geek power user feature.

Someone might buy the company to offer as part of a home automation type product, but that's about it.


I think you are underestimating the potential of IFTTT to become a sort of protocol for many different things among other home automation and for others to build apps on.


Dittos.

And generally my #1 rant about any given tool / webservice / product / etc.

Tell me. In language a five year old would understand. What the fuck your gizmo does for me. What problem does it solve / address?

Prominently on your homepage / project page / blog / manpage / help/about page. Whatever.

That is all.


maybe having the learn more content just be below the tagline instead of having to click a link. friction.


It is a bummer that the Belkin WeMo's are going to be so expensive.

http://www.belkin.com/wemo/ ($50 switch, $100 for switch + motion sensor)

http://ifttt.com/wemo_switch

If I was looking to dip my toe into some home automation toys, I would rather pick up a Nest($250) and one of the iOS/Android powered home locks (~$300) than the equivalent # of WeMos(at $50/per)... but that's just me. To get real value out of a WeMo system, I figure you need at least 10 spread around your house.

If the WeMo's were $20-25/ apiece and they had a wall switch version, I think Belkin & IFTTT would be a game changer in the home automation space.


>one of the iOS/Android powered home locks (~$300)

Check out Lockitron! (https://lockitron.com) - we come with an API, no monthly fee, and a variety of locks that will work on most doors or apartment buzzer systems.


What's the solution when your phone is stolen/offline/out of order?


Lockitron also offers locks with a keypad, and every lock comes with a Bluetooth keyfob for emergencies like this. <3 lockitron.


I don't see any keypads on the site. Is this true?

A keypad backup is a very desirable feature to me. Possibly to lots of others too, I would think.


Ah, yes, a bit of a secret those. We're currently doing field tests with our keypad locks.


$50 for the switch isn't that expensive given (a) it's WiFi controllable; (b) there's not much in the way of alternatives; and, (c) it's consumer focussed.

Prior to the WeMo your best option for power socket control (as an amateur/DIY/Maker) was something like one of these:

* http://www.sparkfun.com/products/10747

* http://www.pwrusb.com/order/

Or hacking a radio remote-controlled power socket. All of which require additional work.

I'm waiting for a 240 Volt version with NZ connectors.


So we need a RasPi with either a $2 solid state relay (on/off) or $8 triac (on/off/dim) and a powerline ethernet interface.

Then its about $45 and open source. Of course it isn't in the fancy package.

Of course if we want to go that route, then an open source board using say an ARM Cortex M4 (like the ST Micro 32F4) would get the parts cost down around $10 - $15. I'm guessing that is how Belkin implements this as $50 is about 3x that cost.


Using a Raspberry Pi is massive overkill. A $20 XBee module, a $3 microcontroller, and a custom PCB will do the trick; you can then control it all from your PC with another XBee radio. It'll probably be easier than making Linux deal with your relay, because microcontrollers are easy mode. Even an Arduino would be expensive overkill for this project.


Once you add a power supply for your micro, connectors, and an enclosure, you'll probably be hitting the $50 mark, if not more.

You also need to be very, very careful when working with the mains. I wouldn't route a board that works with it unless you know exactly what you're doing.

    It'll probably be easier than making Linux deal with your relay
At least on the Beagleboard, GPIOs show up as files that you can write 'high' and 'low' to, making it easy to switch relays.


How is it is massive overkill? The total amount of structure is the same in both of the solutions. Only in the one you mention it has less gates. Same amount of input resources. Gates a free. 5 years from now kids will be making LEDs blink using vixie cron and a python script talking the GPIO lines of a 32 microprocessor and this is O.K.


How is it is massive overkill?

A BCM2835 (Raspberry) uses ~2W, while an ATmega186 (Arduino Duemilanove) uses ~0.027W. So yes, massive overkill.


I think you mean the ATMega168 and its power consumption is somewhat less than 0.0027W, while the BCM2835 is more like 2.5W, but regardless, I think you're defining overkill in terms of power, which is only one dimension, probably not the one the comment was referring to.

It's a bit like saying using a tungsten lamp is massive overkill in lighting a room.


Also, I'm concerned about the idea of it taking power even while everything in the house is off. Is there a page somewhere that addresses the power usage concern?


I tested a preview unit. It was something like 0.3 watts, if I recall correctly.


Perhaps if they added integrated batteries that only initiated a power draw to charge when they reached a specific level they wouldn't need to draw power from the grid constantly. Expensive way to solve the problem though, perhaps.


That would use more power because of losses in charging the batteries and standby loss in the battery themselves.

They could optimize for lower power. 0.3W seems pretty high, I would bet they could reduce that by a factor of 10 if they tried


Can you still send 802.11b/g/n with 0.03W power consumption?


These guys are doing it using just 0.00015W

http://supermechanical.tumblr.com/post/25189115532/its-all-o...

Of course it depends on how often you "call home" to check for any information. What might be acceptable for monitoring temperature would not work for a switch when you expect immediate response.


It might have been 0.03W. I’ll need to check again… It was either 0.3 or 0.03.


I've skimmed the first two screen-heights of that page, and I still have no idea what it actually is or does. Maybe link to an explanation somewhere?


http://ifttt.com/wtf

Edit: Sorry for the trouble. We're working getting a link to our About page from the blog. If you hit the home page, the giant 'Learn more' link has been really successful at getting people there.


Wow, how have I not heard of WeMo before? Though Belkin is typically a little overpriced, I think they've opened the door for other companies (Phillips, maybe?) to offer their own version. I'm definitely looking forward to automating my dorm next year.


I wondered the same thing as I've been researching controllable power sockets for a while (although more looking at digital IO/USB control).

I only discovered WeMo today via a comment on Twitter. Though they were announced back in January as available in "US summer" it seems they've been pretty much under the radar since then until now.


My first thought on seeing the name was that it had something to do with Fourier transforms.


Hell yeah for marketing this is cool:

IF retweeted

Then Turn Strobe Light On

Then Wait 5 seconds

Then Turn Strobe Light Off


Sounds like it could give you a seizure if your tweet went viral.


Ha! It could be a lot of fun to think of potential dashboard (1 for each sale/sign-up seems like a good metric for some startups) related hardware/software interfaces.


I would use IFTT if I could pay them money.

I am wary of building on top of any third party web service, but I free services like this I may play with but never depend on for personal or business use.


I've upvoted you, but I really want to second this verbally too. I don't see any adverts on IFTTT (and even if they were: once setup, I wouldn't be visiting the site often and seeing them), and don't understand how they're monetising. This is a service I'd really like to use, but I don't want to come to depend upon it, then find it folding (either due to monetary problems, or acquired by a third party and being shut down).

Some information on how they intend to stay alive would be fantastic.


I wondered about that too, and one thing that occurred to me was that if they get popular, they've got access to combine data about their users across all the source and sink services the users use. That is, they can get all my readily available Facebook data (friends, likes, etc.), my Twitter feed and follower/followee/favorites lists, etc.

I doubt IFTTT is doing such data fusion today (their userbase is probably too small to make it worthwhile), but I wonder if that's the eventual plan. To advertisers, that sort of data could be immensely valuable for ad targetting.


My first reaction was that businesses would use this to do systems integration (send an email every commit, add a note in group chat before a meeting, etc.) These are already pretty easy to wire, but there's value in going from easy to trivial...


Maybe through 3rd party API licencing like the Belkin (and forthcoming Pebble watch)?


I am happy about their foray into the physical world but there is a much more critical feature still missing in my opinion: the possibility to check for a desired trigger (e.g. stocks) at a specific time/date or, in general, the possibility to combine multiple triggers at once


I very much like it, two comments:

There doesn't seem to be any interface for adding multiple channels of the same type. My use case: a personal gmail account and a google apps email account. In general, it would be good to be able to add multiples of any channel, eg

- multiple time channels for different time zones

- multiple twitter accounts

- multiple sms accounts

The ability to share recipes is great, however it is too hard to browse them. Of the ~35000 shared recipes I suspect many are variations on a theme, and whilst the ability to filter by channel and the ranking by popularity are useful much more could be done. In particular,

- being able to choose existing tags when creating new recipes

- being able to browse existing tags

- suggesting recipes that are similar to the one I am adding, or even creating!

Love to see how this progresses, especially as the API opens up.


I'm not clear on the difference between a trigger and a quick trigger. Is it just that regular triggers are run in batch jobs, and quick triggers are processed individually?


Most of the triggers on IFTTT require polling to receive updates. We poll the sources roughly every 15 minutes.

Quick triggers, on the other hand, are event-driven.

So, for instance, if you had a task that was "IF I send IFTTT a Google Chat THEN post to Twitter" we'll now post your tweet within a couple of seconds (vs. before it would have been within 15 minutes)


Regular triggers are checked every 15 minutes. http://ifttt.com/wtf


This is the first time I hear about IFTTT. I like it very much. The name is terrible though. The other similar startup on HN, Zapier, has a much better, easier to remember name. I'd go as far as to suggest to change the name. Here's one: apibit.com is still available.


"If this then that" is a terrible and hard to remember name?

IFTTT is just the acronym.


If you are looking for 'microcontrollers <-> webservices with full restful api' then Cosm looks pretty good. Arduino, Nanode, mbed, devices running OpenWRT...

They just changed names from Pachube.

https://cosm.com/


The Belkin WeMo looks useful, but looking over its site quickly, it doesn't appear to be programmable through an API.

Does anyone know of a similar product which can be programmed through an API?


it sounds like IFTTT is its API.


Not really excited. There are not many new channels, and those are mainly for some additional devices. Some may find them useful though.


The key thing for the new channels is that it's the first expansion outside the screen and into the physical world. A lot more to come in that area, stay tuned.


like the pebble yea?


Yep!


I'd be a _lot_ more excited by the update if I could add my own custom channels for less popular sites, or use other custom channels that other people have added.


We're working on a Channels API to open it up to devs.


If its an introduction, how can I read the entire post without knowing what the hell it is?


Is there any updates on the ifttt api? When is the ETA?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: