OneTok provides developers with a simple and flexible infrastructure for building voice recognition controls into mobile apps.
The solution is cross-platform, offering built-in support for iOS, Android, BlackBerry 10 and Windows Mobile. Additionally, OneTok provides libraries for integrating voice into third-party app development frameworks, including Appcelerator Titanium, PhoneGap and other HTML5 environments. The platform leverages the phone’s existing hardware and software, and performs audio processing and analysis in the cloud.
The platform offers pre-defined language recognition kits for a series of app categories, including banking, weather and travel. Each category contains a variety of actions and corresponding sample sentences that allow for the automatic set-up of basic voice controls. Eventually, developers will be able to create additional phrases which can be defined and mapped to functions in the app.
OneTok is free to set up and integrate into a working app. An online sandbox offers a fully featured environment to test functionality. Once content, a developer can deploy a OneTok powered, voice-enabled app for free. Developers will then be billed on a per usage basis.
"High Street Bank" integrates OneTok into their iPhone, Android and Blackberry app for customers to benefit from in-app voice commands. The app provides users with a variety of features including voice login/logout, paying bills, locating ATMs and checking account balances.
When the user says, for example, "Where is the nearest ATM" , the app returns a map of the user's location with the nearest ATMs clearly marked. What's great is that the app is able to understand multiple versions of the same command. For instance:
"I want to draw money"
"Show me the nearest cash machines"
"I need to get some cash"
Each of these sentences is stored in OneTok's platform and subsequently broken down and tagged by their respective parts of speech (e.g. preposition, noun, verb, etc.). Based on the context of each word within the sentence and the pre-defined rules that the app developer has established, OneTok is able to match these tags to a pattern that is associated with a specific "findAtm" action. This action is then passed back to the app, where it performs the desired function.
Below is the home screen of Onetok.com.
From Onetok.com users can sign-up to create a developer account.
Account sign-ups will ask for:
1. Email Address
2. First Name
3. Last Name
6. Phone Number
A confirmation email will then be sent to the user in order to proceed with account creation.
Once a OneTok account has been set-up and the user is logged in, an App Summary Page will provide an overview of their account.
From the App Summary Page users can:
1. Add an App
2. Edit a pre-built app
3. View analytics on voice commands received for a deployed app
Upon creating a new app, the user will be asked for:
1. App Name
2. App Category, which will associate the app with a selection of pre-defined actions and sample sentences (e.g. for the banking category this will include actions such as getBalance, findAtm, findBranch, showTransactions and transferFunds and all of their corresponding sample sentences).
Upon creating a new app, OneTok will generate an App ID and App Token that the developer will use to uniquely identify their app when interacting with the OneTok API.
OneTok is able to optimize in-app voice controls by performing targeted audio processing on the client side, stripping out noise before analysis takes place. Each App ID is associated with a set of predefined actions and categories and thus the recognition engine is set-up to expect and seek-out specific words from each command.