Verbal - Voice Internet
Verbal - Voice Internet
Verbal is a new platform that will allow people to use many internet services through their voice using any kind of phone.
Verbal is a new platform that will allow people to use many internet services through their voice using any kind of phone. Read more
Thanks to Verbal, millions of people that are currently not connected could get the benefits of the Internet using their voice and a regular phone line. Help us achieve this vision by gifting them a one-year subscription to Verbal and, in exchange, getting a cool symbol (T-shirt or USB wristband) so you can show the world that you helped us in our effort to bridge the digital divide. Below, you will find an overview of the main rewards, but if you want to try Verbal yourself or develop voice apps for our open platform, please make sure to check the complete list of rewards (on the right of this page).
Scroll all the way down to find a more detailed view of the USB wristband and the T-shirts.
The digital divide remains a huge global issue -- especially in developing countries where only 6% of the population enjoy broadband access in their homes and only 19% in their mobile phones (source: ITU). This means that, overall, only 38% of the world population use the Internet.
Telecommunications infrastructure is a major cause for this problem. Nonetheless, the lack of computing devices and of basic abilities for the use of information technologies among those affected are just as important.
In these conditions, billions of humans around the world fall behind the rest of the population with regards to the economic, social and cultural benefits that the Internet affords.
There are some amazing projects out there tackling the digital divide. However, most of them focus on delivering new infrastructure, computers, or digital literacy programs. As such, they are hindered by steep curves of penetration, adoption and learning.
Wouldn’t it be possible to let people enjoy the benefits of the Internet more quickly and easily?
In contrast to the meager numbers in broadband penetration, the amount of people that have access to voice telephony is staggering: 29% percent for landlines and a whopping 96% for mobile. And that doesn’t even include public payphones!
In addition to that, recent advances in cloud computing and telephony, as well as in automatic speech recognition (ASR), open the possibility of delivering Internet access to billions around the world, using resources that a great majority already enjoys: a telephone and their voice!
This is what Verbal aims to achieve: to deliver internet services and content, through voice telephony, to potentially millions of users in many countries.
How does it work?
When a user dials Verbal the call is connected to one of our Amazon EC2 instances running an Asterisk communications server. These take care of all the computing processes, so that we spare for the user the need for a computing device.
Then, using speech recognition technology -- we are currently working with CMU Sphinx, Wit.ai and Google's Web Speech API -- we convert the user's spoken words to text. From there, it works just like the regular internet: we can use that text to perform a search, send an e-mail, etc.
Finally, if the service returns a text (like a search result or an e-mail response) Verbal simply speaks it back to the user.
We started out building a proof of concept using off-the-shelf cloud telephony solutions (i.e. Tropo), which included information search using Wikipedia and an e-mail service using hard-coded addresses. Nonetheless, we quickly realized that our project would require a full ad-hoc development. That's why we decided to reach out for help on Kickstarter!
At this point, we have a functional prototype, in English language, of an information search service using Wikipedia. For this, we set-up our own cloud telephony system using Amazon Web Services and Asterisk, which we can use in conjunction with either Sphinx or Google Speech API for speech-to-text conversion. However, we still have a lot of work to do.
With your support, we will be able to complete the development of Verbal in the following areas:
1. ASR engine: The ASR engines currently available are optimized for English language and mostly require audio in 16khz. Nonetheless, our project require engines that can handle 8khz audio, used by regular telephone lines. Also, we want to bring the benefits of Verbal beyond the English-speaking world, particularly Latin America. Thus, we need to optimize our ASR for Spanish language and narrowband, 8khz telephone audio. This part of the development will take about 40% of the funds we are asking for.
2. Basic suite of voice apps:
- Search: an information source connected to services such as Google, Wikipedia, Wolfram Alpha, etc.
- E-mail: a messaging service for delivering transcribed voice messages to inboxes and reading e-mails to Verbal users.
The development of these apps will take about 20% of the total funds we require.
3. Developer platform: A software development kit for developers and the platform required to publish their apps, which will take about 20% of the funds.
4. Launch: deals and set-up with telephony carriers in USA and Mexico, which will take the final 20% of the funds.
The project will heavily rely on the power of it's community. That's why, all the code behind the Verbal engine will be accessible for developers in their way of finding solutions for the people.
An entire SDK with it's proper API will be fully documented before Verbal's international launch.
Sebastián Tonda - CEO & Partner, Flock
Sebastian is a Mexican entrepreneur. He is the founder and CEO of Flock, one of the leading digital marketing agencies in Mexico City, with offices in Chile and Argentina, and working for clients such as Telmex, Coca-Cola, and Nike, among others. He is also co-founder of Swarm, the first Latin American research agency specializing in digital market research. He has bachelor's degree in Communications from the Universidad Iberoamericana (Mexico) as well as two Executive Program Certificates from Singularity University (USA).
Sebastian is also ambassador in Mexico of Singularity University, an non-for-profit organization based in Silicon Valley whose aim is to educate, inspire and empower people to face humanity's great challenges that through knowledge regarding the exponential development of technology. Sebastian is primarily interested in studying how technological milestones will shape humanity. For Verbal, Sebastian is responsible for business and strategic development.
Mario NIssan - CIO & Partner, Flock
Mario is a Mexican startup investor and marketing technologist. As partner and CIO of Flock, he is in charge of overseeing technology strategy, IT operations, performance, as well as Big Data and analytics. He is also a board member of Swarm, which uses semantic analysis technology for market research. He is an investor and mentor for startups such as Zuumd and Clinch Labs. He has a B.Sc. in Industrial and Systems Engineering from the Monterrey Institute of Technology and Higher Education (Mexico) and an Executive Program Certificate from Singularity University (USA).
As a former CEO of Tecnosinergia, a company specialized in security and wireless infrastructure, Mario has ample knowledge and experience in the technology and business strategies of the telecommunications industry. He is passionate about how technology is redefining the way we live and enhancing the overall human experience. For Verbal, he is responsible for infrastructure and business implementation.
Jorge Camacho - Director of Innovation, Flock
Jorge is a Mexican thinkermaker. As Director of Innovation at Flock, he is in charge of special projects involving creative technology as well as product and service design. He has a Bachelor's degree in Communications from the Universidad Iberoamericana (Mexico), a Postgraduate Certificate in Research, an MA in Cybernetic Culture and a PhD in Cultural and Innovation Studies, all of them from the University of East London (UK) and, recently, a Certificate in Innovation and Design Thinking from the Centro de Estudios Superiores de Diseño de Monterrey (Mexico). He has lectured on communications and technology, interactive media and design research at the Universidad Iberoamericana (Mexico) and CENTRO (Mexico)
His main interests, both academic and professional, are the empowering of people through new technologies and the development of those technologies using human-centered design methods. Thus, for Verbal, he is serving as project director and responsible for user research and development.
Mauricio Ruiz - DIrector of Creative Technology, Flock
Mauricio is a Mexican hacker and entrepreneur. WIth only 21 years, he leads the Creative Technology department at Flock, where he infuses the startup genome to every project. Building fast and failing cheap are his mantras. He's finishing his B.Sc in Industrial and Mechanical Engineering at UNAM (Mexico) with a research scholarship from Carso Foundation (Telmex, the largest telecommunications provider in Latin America). He has a strong background in software and hardware development, using Objective-C, Ruby, Python, C#, Processing, and C as its main tools. Speech and sound technologies are his current passions.
Most of all, Mauricio is interested in using technology for the good of people or just to make the world a little bit easier. He aspires to become an applied physics professor, to put Mexico on the spotlight with his technology and social innovation projects and to trigger an urge to start up in whoever listens to him. For Verbal, he is responsible for technical development -- he is the man behind the magic.
Of course, Verbal is an effort that extends well beyond the core team. Thus, we want to acknowledge the importance of Flock's leading creative team -- Daniel Granatta, Eduardo Salles and Nicko Nogués -- whose input has been, and will continue to be, instrumental in the development of Verbal. Also, Carmen Ruiz's strategic thinking was central in the early stages of the project.
We want to thank also our fellow 'Flockers' Virginia Irizar, Alejo Sarquisse and Raúl Islas for their excellent work on the project's video; Alejandro Ramos and Felipe Carral for their valuable opinions and feedback on the project; Juan Carlos Tonda for his help and to everyone at Flock for their support.
Risks and challenges
We have detected two main challenges. On the one hand, the further development of the speech recognition engine is not a small feat. Specially considering that we are developing it for English and Spanish languages, both of them for narrowband (telephone, 8Khz) use. Nonetheless, we have gathered a great team of developers and we are making great progress. So, we feel pretty confident about achieving the feat.
On the other hand, the implementation of Verbal requires deals with fixed and mobile telephony carriers. For Mexico and the USA, we are already quite familiar with the required deals and have even began negotiations. Nonetheless, the expansion of Verbal beyond these countries will involve a lot of negotiation with national and local carriers.
- (51 days)