Alexa Skills Kit
The Alexa Skills Kit is a software development kit (SDK) that enables a developer to build skills, also called conversational applications, on the Amazon Alexa artificial intelligence assistant.
The Alexa Skills Kit is comprised of tools, application program interfaces (APIs), code samples and documentation that enables a developer to add skills to the 10,000-plus voice recognition capabilities available on Alexa.
Amazon Alexa is based in the Amazon Web Services (AWS) public cloud. A developer can upload Alexa skill code to AWS Lambda functions to execute code that is triggered by voice interactions. AWS automatically manages the compute resources for Lambda. A developer can certify, publish and update skills, which are made available through the Alexa Skills Store.
An organization can build an Alexa skill to connect to end users via the conversational Amazon Echo platform. A developer programs the voice user interface to return a variety of voices, accents and responses based on the code for the skill.
Three command-oriented skill options
A developer has the ability to build three types of Alexa skills: custom skills, smart home skills and flash briefing skills.
Custom skills allow the developer to define requests that the skill can handle, define the vocabulary required from the end user to make the skill request, and define the invocation name for Alexa to recognize the task. Custom skills are flexible and can handle several requests, but they require the most code to generate.
Smart home skills control cloud-enabled devices, such as the ability to control lights and thermostats automatically, in a smart home. The Smart Home Skills API uses Amazon's standard language processing model. A developer does not need to build a custom voice interaction model with smart home skills, but must create code to respond to a request.
Flash briefing skills enable businesses to share pre-recorded audio content from feeds on Alexa with end users, such as news and weather updates. The Flash Briefing Skill API uses Amazon's language model and removes the need for a developer to build a custom voice interaction model.
