Google has created a framework for building an AI agent. This framework is called as a Google Agent development Kit. And it also has variety of the language SDK for the framework.
I recently covered the Typescript SDK and now I also covered the Golang SDK with the Google ADK. I made a video for you to check out the whole process.
In this particular Agent I decided to turn the framework output into the CLI agent. It made my life easier when I used this SDK and decided to create command line output to build the final agent.
In such agent development it helps if you sign up to a service where you can get the free AI inference API. Like say Kilocode or Openrouter or similar other type of the service. I also think you can use Opencode with their free models in the API.
In this tutorial I decided to use the Weather API and that API I managed to get access to certain APIs endpoint and some of the city data for the weather related output.
At the end I made this agent such that it answers most of the weather answers, and for it makes use of the LLM output where the weather API data ends. So both those options kind of contribute for the development.
I'd recommend you to check out the Google ADK and use any SDK of your choice - typescript, python and golang SDK. I feel Python and golang just are too good and quick to build the command line apps.
Top comments (0)