Google now lets hardware developers tinker with its Assistant

Google has long said that it wants to open up its Assistant to a broader ecosystem of hardware companies and developers. Today it’s taking a major step in this direction with the launch of the Google Assistant SDK, which allows developers to build the Assistant’s smarts into their own hardware prototypes. Those could be smart mirrors, Google Home-like appliances or a teetotaler’s favorite robot bartender — the world is your oyster.

It’s worth stressing that Google is only making the API freely available for building prototypes for now. If you want to commercialize a project, you still need written permission from Google.

The Python-based SDK works with a number of hardware platforms, including the popular Raspberry Pi 3. Using Google’s gRPC API, developers also can integrate the Assistant into other hardware platforms and use languages like Java, C#, Node.js and Ruby. The SDK makes it easy for developers to listen for spoken commands, send those on to the Google Assistant service and then receive the appropriate response.

Google says that devices that use the SDK will be able to use the full power of the Assistant, but that’s not quite the case today. According to the release notes, you can’t set alarms and timers, for example. Neither can you play music, news or podcasts just yet. And to enable most home automation features or user other third-party services (like using the Assistant to call an Uber), you still need to own an actual Google Home.

It’s worth noting that Google already offered an API for talking to the Assistant, too. This “Actions on Google” API, though, focuses on integrating third-party applications with the Google Home devices (with support for the Assistant on the Pixel phones or in Allo coming later).