The SDK enables developers to create devices that allow users to talk to them and get questions answered. Still in a preview form, the kit includes a gRPC API, a Python open source client to handle authentication and access to the API, samples, and documentation. There are gRPC API and bindings for languages like Java, Python, C#, Node.js, and Ruby.
"The SDK allows you to capture a spoken query, for example 'what's on my calendar?', pass that up to the Google Assistant service, and receive an audio response," according to a bulletin posted by Google. "And while it's ideal for prototyping on Raspberry Pi devices, it also adds support for many other platforms."
Features still remain to be built, however, including hot-word support and companion app integration. "We've created a new developer community on Google+ at g.co/assistantsdkdev for developers to keep up to date and discuss ideas," Google said. "There is also a
stackoverflow tag [google-assistant-sdk] for questions, and a mailing list to keep up to date on SDK news.”
Google Assistant brings capabilities like natural language understanding and voice control to devices. The technology was made available for some Android smartphones earlier this year.