Twilio does NOT do speech recognition. Their Gather verb collects DTMF touch tone inputs. They do not need the complexity of VoiceXML because they aren’t doing complex ASR interactions.

Angel and TellMe were mentioned above as comparisons. They both do speech recognition.

Twilio does NOT do speech recognition. Their verb collects DTMF touchtone inputs. They do no need the complexity of VoiceXML because they aren’t doing complex ASR interactions.

Anyhow, I have never understood the VC funding model coz some of the investment they make just beats my simple business logic.

For the sake of Twilio founders, I hope I am wrong and I hope Twilio is right.

I doubt Tropo.com will have a negative impact on Twilio. The market of telephony mashups is just starting to develop and it is bigger than one or two players.

Telecoms are also starting to offer similar developer friendly APIs. I am referring to BT and Ribbit.

A quote from the Telco 2.0 blog about the World Mobile Congress last month sums up the trend:
“Who would have imagined, not so long ago, that none other than AT&T, the biggest ugliest telcoest telco in the world, would be talking about the importance of ”standards-based APIs that ensure consistency, security and ease-of-use for developers”?”
http://www.telc...t/blog/2009/02/

For full disclosure, I work at Ringful, which plays in the space of Mobility Mashups, offering a RESTful Web Services API to a rich set of telephony functions with worldwide access.

Did today’s announcement by Voxeo just kill Twillio? eek.

BTW, I <3 ShoutNow

I recently started a dating website hintcafe.com and built some interesting features like leaving a voice message for a user and calling another user without knowing their number.

For some details, see http://hintcafe.com/faq. For these things in action, see my example profile at http://hintcafe.com/user/vivek

I built the following on HintCafe using Twilio

a) Phone validation – users can add their phone number to the account. The phone number is validated using something similar to Twilio add-a-caller-id process. User is given a 4 digit validation code on hintcafe.com and user enters that code when hintcafe.com makes an outgoing call to the phone using Twilio make call api. The validation code is captured using Gather.

b) Leave a voice message – The user calls hintcafe.com and then user is asked to record the message for the chosen user. This is done using Record verb. The message receiver is sent a flash player linked to mp3 of their recording on hintcafe.com inbox. The recording mp3 is fetched using recordingurl provided by Twilio.

c) Call Forward – Users of hintcafe.com can talk to each other without sharing their phone number. The caller calls Hintcafe.com incoming number hosted on Twilio. HintCafe uses a dial to call the other user using callerID of HintCafe to make the dial. This way, both users can talk to each other without sharing their phone numbers.

My last week developing the solution with Twilio has been a very positive experience. The Twilio team on help@twilio.com and here on forums is very prompt and concise in reply. The overall product is well designed with the REST APIs and is simple to user. And ofcourse, the pricing model is just fantastic which makes it capable for early stage startups like mine to build interesting features without considerable upfront cost.

They support http request style control over their interface.

Twilio’s philosophy is to make the world of telephony accessible to any web developer with a familiar pay-as-you-go pricing model. That means letting developers build high quality voice apps using PHP/Python/Ruby/Java/C#/ASP or whatever language they want, let them host and improve the code on their servers along side their existing web code, and let them reuse existing tools, development processes, and database logic.

The founders of Twilio are all developers and we’ve tried really hard to provide the documentation, sample code, and community support network to let developers build awesome applications. We’re so excited by what developers have already built using Twilio and can’t wait to see what’s next. If you haven’t seen our new homepage we highlight a range of excellent new applications built on top of Twilio and the developers who built them.

Twilio is just getting started. We have a giant list of awesome ideas and we are going to be continually rolling out new features over the coming months. In fact, look for an announcement later this week

Cheers from Twilio,

-Evan

(if you think it looks like a button, you should keep your zipper shut)

Trying to reinvent the VXML wheel will not go anywhere. VXML is very straight forward javascript based language. Learning curve is 1-2 days on a maximum front.

Ribbit is more flash based phone rather than phone logic controller.

This is nice functionality, but there are numerous people who have come at this problem from similar perspectives.

Here’s what’s right: telephony apps will be delivered via the cloud, and they will be assembled from pre-built modules on someone’s app server.

Here’s what’s wrong: this is still a telco play, meaning the company is in the minutes business. Margins will approach fractions of a cent a minute, not the $0.05 cent level they’re hoping to get now. So, fingers crossed these guys can get people to deliver billions of minutes across their platform.

The comps for this are not TellMe. Ribbit is fair comparison, but this is really something Telera was doing almost 10 yrs ago. Angel is probably the best comp.

No way the investors understand what they put money into.

I think the designer must have been getting dressed when he got the inspiration for this one.