At TechCrunch Disrupt Europe in Berlin today, Voicesphere is launching a platform for voice-enabling consumer and enterprise apps and making them available in a Siri-like environment based on natural language processing.
The service was founded by David Kokkelink and Jan Thöne, two recent high school graduates from Germany. It features both a consumer-focused app built by Voicesphere itself and – maybe even more importantly – an API for enterprise developers who want to enable their apps for use within the service. Ideally, the service wants to become your main hub for interacting with your phone and all the services you use online.
As the company’s founders argue, Siri and services like Maluuba can only interact with a very limited amount of services. Ask them a question and unless one of the connected services can provide the answer, they’ll be left speechless. No single company, Voicesphere believes, can comprehensively cover the wide range of information users are interested in, so the only way to approach this problem is to build a more open system.
For now, Voicesphere isn’t fully open. It’s focusing the developer part of its platform on enterprise solutions, and it’s also building a consumer-facing app, which integrates services like Instagram, Facebook, SoundCloud and others (for Disrupt, the team also integrated TechCrunch’s RSS feed).
Enterprise developers will soon be able to integrate their own services into Voicesphere. They will be able to take their own web-based APIs and use the service’s browser-based development environment to create so-called “voiceapps” that make their services available to Voicesphere users.
As the company’s founders told me, the idea here is to create an enterprise app store in Voicesphere and to allow companies like Salesforce or Microsoft to charge users for using their apps using the company’s voice platform. Voicesphere will take a 20 percent cut from these sales.
The team argues that this frees these companies from having to learn about natural language processing and allows them to focus on building the best app possible while still making it available through voice commands in Voicesphere.
The IDE for enterprise developers, the team tells me, will launch within the next few months.
Voicesphere For Consumers
Enterprise services, however, are just one aspect of the service. What’s launching today is the Voicesphere consumer app. It supports about 20 consumer services – most of which don’t feature any voice recognition and natural language processing features by themselves. Indeed, because it just uses the web-based APIs of these services, users don’t have to have any of these apps installed on their phones.
This system, the team tells me, is very flexible. Users can speak in conversational language to, for example, ask the system what’s new on their Facebook feed (you just ask “what’s new on Facebook”). They can then update their status, too. Given the potential for embarrassment (voice recognition isn’t perfect after all), Voicesphere then double-checks and asks if it correctly understood that the user really wants to post a Facebook status update. Other use cases include liking images on Instagram by voice as well as finding songs on SoundCloud, searching through your files on Dropbox and interacting with other popular consumer services.
The app also features a basic news feed that aggregates updates from all of the services the user has activated in the app. Tapping on one of these then allows you to start interacting with the update by voice.
In the long run, the team will likely open this part of the service up to developers, as well. While the IDE will first only be available for enterprise companies, the team told me that it will likely open it up to developers of consumer apps as well.
To be successful, Voicesphere will have to attract enough developers to support it and users who will install it. So far the founders haven’t reached out to any enterprise developers yet, so over the next few months, they plan to research this market more deeply and then launch the service with a couple of partners.
The Voicesphere consumer app is now available for download on Google Play.
Q: What are the limitations? What apps can you integrate?
A: Any app with an API.
Q: What technology are you using for the voice recognition.
A: We are using Google’s technology right now. Google’s API is the best on the market today. We focus on providing the natural language processing and the API infrastructure.
Q: How do you make money?
A: We will sell enterprise integrations for a fee. We want to open up our API Hub to professionals and developers who will be able to sell their integrations.
Q: What would you do with funding?
A: Expand the number of integrations and look at additional platforms.
At the end of the presentation, AngelPad’s Thomas Korte also invited the team to apply to AngelPad in New York.