Every once in a while we come across a company that seems to have a giant bullseye on it for acquisition, with a great product, viable business model, and a talented team. Twilio, a startup that has created an intuitive API for a variety of telephony services, is that kind of company (it also managed to Rick Roll my boss). The startup has developed a simple API with pay-as-you-go pricing that allows developers to quickly implement phone services into their applications, opening the door to a number of services that were previously only accessible to the small sliver of engineers trained in the dark magic of phone calls. Twilio is launching today in private beta, and TechCrunch readers can grab an invite here.
CEO Jeff Lawson says that while other web telephony services exist (like Asterisk, an open source project), these technologies tend to be very complex and difficult to use, even for experienced developers. Lawson says that Twilio is looking to commoditize these phone services by making them much more accessible to developers, by introducing a set of very intuitive commands. The API primarily consists of 5 commonly used phone actions (Say, Play, Record, Dial, and Gather a phone number), each of which behaves exactly as you’d expect it to. That Rick Roll app we heard a few days ago? Here’s the code (for you non-programmers, this is pretty basic stuff):
Lawson showed me a number of other impressive examples, including a project that he said managed to replicate GrandCentral’s core functionality in only around 15 lines of code. A number of organizations have already started using the API to build their own applications, including a non-profit that has now automated hundreds of calls that used to take staff hours to make.
Twilio is adopting the cloud-service model, with no contract required and flat fees for calls depending on the number of minutes used and the number of phone numbers needed (developers can also scale their needs based on demand, so they don’t have to worry about their servers crashing). And while the Rick Roll app was created with the service, Lawson says it was just a pre-launch joke, and that safeguards are in place to prevent any future applications from making annoying phone calls.
Twilio isn’t perfect – it doesn’t yet support voice recognition, which is a key component in many telephony services (though this feature will be released in a future version). But it is very cool, and will probably be very popular among developers. Don’t be surprised if this one gets snatched up soon by a cloud service provider like Rackspace or Amazon (my money’s on Amazon – CEO Jeff Lawson was a Product Manager for AWS).
There are a few of other startups trying to make phone services more accessible to developers, including Skydeck, which we covered here.
And after being featured on TechCrunch I’m sure that it will get acquired.
Who is stronger behind the curtains, Ribbit or Twilio? It seems like they each have simple APIs for VOIP developers. Not sure about pricing or scalability of either. Can they do millions, tens of millions, billions of calls?
Capacity? Speed? Redundancy? Detection? These are the questions that should be considered before valuating this service. The features and API are great…in theory.
So what’s the invite code? None of my variants of “TechCrunch” worked. Am I daft?
I’m very impressed, this service is well thought out and everything you need is there to get you going. I see a huge potential for my company here to be able to build customer service applications, sales applications, etc.
congrats Jeff & Twilio team on the launch!
(and sorry about the RickRoll mike, but I couldn’t resist
I am daft, you just need to request an invite and you get one automatically. /foreheadslap
GREAT ARTICLE! Twilo is a great company. Kudos to them!
If you like Twilo and Grand Central, you should check out a start-up from Silicon Valley called VingTalk:
http://www.vingtalk.com/cpp
Don’t market your crap alternative here. Spammer.
I had the pleasure to meet Jeff at Startonomics — it was immediately obvious to me how valuable and powerful Twilio was when he told me about it.
It’s great to see credit where credit is due. They’ve really built a massively useful platform.
Congrats Jeff on the great review.
Pardon me, but what exactly is something like this useful for?
Curious George,
I can tell you exactly how this was useful for my company (http://www.life360.com). We are building voice-intensive emergency-focused telephony applications, and the last thing we want to do is deal with port capacity, complex service contracts, new programming languages, or infrastructure maintenance.
Before Twilio we had 2 options: A) go with a lightweight cookie cutter service that offered little flexibility, or B) sign a big contract with one of the large enterprise oriented providers.
For our proof of concept, we went with option A, but quickly outgrew it and had to shift to option B almost immediately. Switching to option B cost about $8,000 upfront, and we were looking at thousands of dollars a month in fixed fees once we went live.
With Twilio we got the best of both worlds—an exceptionally straight forward and simple development platform that is also scalable and cheap (with no fixed fees or complex contracts). If you want to see something tangible, we have a use case for one of our products here: http://life360....-emergency.html. I met with Jeff and John and they surprised us with a working version of the app they coded themselves in 20 minutes.
To me as a startup CEO this is awesome—everyone knows that lack of focus is a killer, and the fact that we just solved our voice problems in days instead of months means we can launch sooner, spend less on coders, and stay concentrated on our core service.
Chris
just curious, what about Ribbit? How similar or not is that to this?
Marin, without going into too much detail, Ribbit is also awesome and we are planning on using both companies for different chunks of our service. For instance, Ribbit can do really cool stuff with VOIP, and we made a flash-based “emergency phone” that lets our users place calls from within the browser by clicking on a person’s icon within a map (so in Katrina where the web was up but phones were down, people could have used our Ribbit-based app to make calls to people out of the area, or even browser-to-browser to other people affected by the same event). Ribbit also has its own consumer product which gives end users their own numbers and is quite different from Twilio.
My guess is both services could step on each other’s toes in the future but for now solve different pain points….maybe Twilio is the next BT acquisition?
With all due respect, it seams relying on a young startup with unproven technology as the backbone for your “voice-intensive emergency-focused telephony applications” might not be the best idea.
I mean this service sounds fantastic and I’m sure lots of cool applications will come out of this, but building a life-or-death emergency based service on top of it is frankly terrifying.
I don’t know a lot about this niche but I’m pretty certain there’s a very good reason the Telco offerings cost an arm and a leg.
These are companies that have been dealing with natural (and man made) disasters and emergency response for decades.
Anyhow, I hope you have good insurance
Let’s say you have dating site and you want to add a feature where users can leave voice mails for each other, on the site, without ever disclosing a phone number. On each users account you have a button that says something like “Message Me”. you click that button and the system calls your phone number (the one that you have registered against your account). You can then leave a voice message for the person you’re interested in. They get an email saying that they have received a voice mail – which they can then listen to on the site. Then they can call you back and leave a message with the click of a “Reply” button. You get an email, etc.
All of this functionality can be coded up in about 2 to 3 hours in Twilio – I know because I’ve done something very similar. It uses your REST services, your app DB – no external code, pay only as you go.
I don’t work for Twilio but I have been an alpha tester for some time.
Nice service, but there’s absolutely nothing new here. There are many large scale companies offering voicexml servers and app hosting, and then there are free voice capture APIs from companies like myvox.com.
Most companies have offered a metered usage model for, oh, about a decade.
But still, it’s just a new way to sell minutes. Do the math. Assume 50% mark-up in minutes, .25-.5 cents for amortized cap costs, and if the do 1B minutes a month (which they won’t) at a net of 1.5 cents a minute, and we’re talking $15K a month in revenue.
Yep, they’ll be lining up to buy this company.
Very disappointing Techcrunch can’t see through this.
There is a very profitable company called Voxeo with 20,000 developers that does just this. And they do it very well.
http://www.voxeo.com
There are a bunch of others including Plum, BeVocal (Nuance), TellMe.
The benefit of using these others is that you dont have to learn TwiML, but rather code in industry standard VoiceXML 2.1.
True, but I think all these other companies have large (for a start-up) upfront costs, and (I’m not 100% here) fairly large per minute usage – correct me if I’m wrong. Twilio has no such up-front costs and their per minute charges are low. It’s more of an Amazon EC2 charging model than a RackSpace model. I’ll give you that it’s not VoiceXML but it’s 5 tags, so if you understand VoiceXML you should be able to get this up and running no problem.
To to open, I’ve known about Twilio for a long time and have integrated it into some apps. In my opinion it is really easy to use.
So the innovation here is the pricing model?
Not sure what is really unique about what these guys are doing. Ever heard of VoiceXML, it’s been around for may years and there are many companies that provide the hosted telephony model (TellMe, Voxeo).
Congrats Jeff on the launch!
I took a look at Twilio and although it looks like it’s a powerful API for routing it’s quite a bit away from GC and Asterisk. Users want to present a number to the outside world and have that routed. We’ve seen these services that call you with weather or send SMS from ur voice (Jott) or literally try to impersonate grandcentral. What makes GC a success is that it does what you want it to out of the box and it’s free. I know this sounds short sited but…end users are also short sited.
Off topic question: what IDE/text editor is that (in the screenshots)?
coda
http://www.panic.com/coda/
We’ve posted at the same time. I had found it when you posted, but thanks anyway.
I’ve found it, it is called Coda. Beautiful interface.
Best profile on techcrunch in months. The voices are a little funky, but unbelievable.
I think its a great product. We have been running Numbr.com (www.numbr.com) successfully (& profitably) for the past 2 years and this gives me a bunch of ideas. Perhaps we should talk. Unfortunately there is no way to connect with the folks at Twilio.
On a side note, acquisition is far far away but I need to understand Gilligan’s (above) math a bit.
1.5 cents * 1B minutes/month = $15K/month huh…
Good luck and wish you the best.
Sorry, typo. I meant 1 million minutes. Assuming they have 1.5 cent margin on each call (very generous), they’d generate $15K revenue on 1M minutes. Tem million mins would give them $150K, etc…
Now a BILLION minutes at 1.5 cents would be really exciting!
Thanks for pointing out the typo. But I still hold to my primary question which is how TC gets excited by something like this.
The real differentiation is in simplicity and abstraction of the telephony world from web developers. I don’t believe this is aimed at people experienced with VoiceXML anyway. This is for everyone else – who can now add rich voice features to web applications. Think about Salesforce.com based applications that send a voicemail or reminder based on lead activity. There is a lot that can be done once telephony is made simple and that’s the real value here.
Just did a quick search. Found a very similar offering voiceshot.com. I have nothing to do with this site but their programming interface description sounds very similar
So, what’s the invite code?!
As mentioned by a few others…
There is nothing new with Twilio. This same concept has been tried by many other companies in the past with million$ rai$ed. Yeah, most of those enabling technologies/softswitches/IP-PBX has been built as a platform and/or software for the enterprise/carrier markets and/or for service providers to build and deliver SaaS / OnDemand Voice for variety of developers in the market place but it hasn’t gone anywhere. I’m sure it will change soon, but zip for now…
Based on the TechCrunch mafia connection, Twilio’s investor/advisor Dave Mcclure pinged / rick rolled Mike Arrington and now this article? C’mon… Why don’t you guys write up on companies that offer hosted voice platforms? They have been doing it for years and have paying customer generating revenues upwards of $50M+
Companies such as Twilio does probably simplify and offer API for those who do not know VXML or proprietary markup languages to control the various call flow/IVR/TTS/ASR stuff from the web to enable Web 2.0 guys to build stuff, but c’mon… Twilio hasn’t proven anything; can they even support 10+ calls per minute or have appropriate reporting, provisioning, monitoring capabilities and etc.???
This article is ReFuckingDiculous and please talk about this guys when they have proven that they can support a decent amount of calls per second and/or even be able to guarantee a decent uptime.
Telecom infrastructure is not a joke and it wasn’t built overnight. There is a reason why meshing voice and the web is so god damn difficult.
Katie I couldn’t agree with you more.
I totally agree with you as well, Katie. This article would have been interesting in 1998. Unfortunately for twilio and their investors there are 7000 other companies doing the same thing and doing it well.
@kia Just hit the request invite and it will spew out a invite code for you. Then go through the sign-up using that code.
It does look like an interesting product offering. The concept of offering it as a platform makes a lot of sense. This could put a damper on places like ringcentral, grandcentral, etc. Would allow anyone to expand this. The big question will be scalability for call volume and essentially having a competing price long term.
Twillo looks really nice, but for now only allows calls whithin the US. Does anybody know a similar service (wrt simplicity) that does allow european calls?
Thanks!
I have created 2 services integrating web and phone – WilluSit.com and WilluWork.com. Both make calls to find people on your phone list that are available to work – one for individuals trying to book one of their Sitters for the night – the other for companies to find a worker to fill a shift. Just a couple clicks, calls are made, when worker accepts they are connected with the person that put out the request via call back.
My point, started these projects over a year ago using IfByPhone.com. The technology has been available and it’s not that challenging to integrate. Not sure if Twilio can match all the functionality but the PRICE is great – less than half per minute charges with no minimum.
I will migrate when my volume increases if I can match functionality.
Thanks,
Paul
Hi Paul,
Drop us a note at help@twilio.com and let us know what your functional needs are. I bet we can help you out.
Regards,
-jeff
Jeff, Dave, Twilio dudes:
do you provide functionality for SMS? For example, I use a number of phones on a daily basis (vonage lines, 2 mobiles, land lines). I’d like to choose which outgoing number is used (for callerid displayed on receiver’s line).
That is pretty much offered by Grandcentral now, and formerly by TalkPlus, although w/ GC you’re stuck w/ using your GC number as the callerid. But the big piece missing is that folks can’t text you back at that number. Of course you’d also need to ask me which mobile num to direct the text to.
Doable w/ your service?
Ya know, I read about a Los Angeles based company on (oddly enough) http://www.kizmeet.com, that was apparently an ad, and I can’t find it now, but they were a 4-man team that has developed this prog that can pull all api and recompile it so it can be used from the core…
Hello TechCrunch readers,
Thanks for all of your awesome feedback and responses here. To those who point out VoiceXML, I’d like to respond. VoiceXML has indeed been an enterprise standard for several years now, and it’s been quite successful for telecom engineers whose job is to build voice applications.
At Twilio, we feel that something a bit simpler, both from the technology and business side, will help a broader audience of developers to engage in building awesome voice apps. Each developer is free to pick the tool that suits their needs and skills, and we think there’s a time and place for all kinds of solutions.
We wish the best to everybody who’s contributed to this discussion, as well as everybody in this industry. It’s an exciting time to be building voice applications.
-jeff
Come on now Jeff. VoiceXML or CallXML doesn’t require you to be a “telecom engineer.” Its pretty straightforward, and if you’re a decent programmer can build a app quickly with any sort of scripting language that dumps out proper XML.
I am a big supporter of Voxeo, as they give developers all the tools needed to play around before they go live. Additionally when you need to go into production they can support you.
Skype Support – they got it
International Numbers – they got it
Toll Free numbers – they got it
text to speech – they got it
voice recognition – they got it
call bridging – they got it
call recording – they got it
Incoming/Outgoing Call Scaling – they got it
What I don’t really get is why Twilio portrayed as being the next best thing by TechCrunch.
I totally agree with what Katie says above.
BTW I have had a production application with Voxeo for more than a year and a half.
Agreed with everything.
But I have had experience working with large banks (Merill, Goldman etc) where you are compared with web21C etc when you are implementing a voice mashup solution when their CRM and Line-Of_Business Applications. The point is we could not have done that implementation with Voxeo.
So, these guys (twilio), jut like Ribbit are trying to create a buzz with students/universities now. Perhaps that is the right thing
Everybody and their mother have a platform and an API. Who has got the real customer? None –
Considering Skype has only being profitable once (when they sold it to ebay) and how VoIP startups all around are dying I don’t see what’s the fuzz with this.
During the VoIP hypefest of years past, it would’ve made an impact, but now?
Shadowlayer
LIgnUp did this and had interfaces to Call Control and Media Control about two years ago. Similar stuff and a lot more. We had something called Codelign which allowed partners to do so.
SG
Anyone know a service that is similar that work in Australia? Or better yet, if these guys plan on making the service available in Australia?