On November 14th natural language processing startup, Maluuba, took the next step in its master plan to enable Android-based device owners to carry out Siri-like interactions with their mobile phones. Its “natural language API” or simply “nAPI” (pronounced nAPI) was made available to approved developers. As documented here, nAPI supports calls from mobile apps using Ruby, Python or Java. It has two flavors: the “Interpret API” can understand the meaning of phrases in selected knowledge domains and the “Normalize API” figures out standard (“normalized”) ways to render common values, such as time of day, dates or currency.

Together, they are designed to make it easy for mobile application developers to give users a mechanism to say or type in instructions or search terms using terms using their own words. As of today (November 20, 2012) nAPI supports over 20 categories of discussion. “General Knowledge” is the broadest category, enabling users to ask questions like “Who was president of the United States in 1948?” or “What’s the tallest mountain?” But, most of the promotional literature focuses on the ability for developers to build apps that can book travel, find a local restaurant, dictate an email or simply initiate a telephone call.

Although it is described as the “Siri competitor for Android,” which implies that it is a virtual personal assistant (VPA), Maluuba is merely a resource to support natural language input for a specific (and growing) list of tasks. That input can be keyed in or spoken, using Google Voice’s speech recognition resources. As for support of spoken output, today the company focuses on using the screen to display results based on natural language input. This, in and of its self, is laudable, and it appears to be making great strides in accustoming Android ownsers to use their own words for input. It just doesn’t happen to be direct competition with Siri or any other application that combines both spoken and visual output as appropriate.

Because it has published a documented API, Maluuba will most likely find that developers add spoken output and other refinements to a wide range of apps. It’s taking an approach that has already attracted $2 million in venture money from Samsung and, according to this report in TechCrunch, it plans to raise another $8-10 million to support its efforts to add more domains, support more developers and add more languages and geographical regions (local data being an important part of international expansion).

Beyond the API, it is unclear whether Maluuba will be adding development tools or accelerators to simplify or reduce the effort it takes for mobile app developers to add voice to their wares. Xtone and Onetok are taking such a the platform-based approach. In all cases, these upstarts (or start-ups) are banking on the fact that Apple is very likely to keep Siri and any APIs it employs, inside the iOS walled garden. Apple is likely to make Siri the conversational interface for iPhones, iPads, iPods, Apple TV and the full range of iOS-based devices (soon to include MacBooks).

Devices “for the rest of us” will benefit from the energy and creativity that the growing community of mobile app developers pour into efforts to add a natural UI to everyday activities and controlling all those non-Apple-branded products for the rest of us.

Categories: Articles

Tags: Speech recognition, Google, Android, Natural Language, personal virtual assistants