Building AI agents with Claude in Google Cloud's Vertex AI | Code w/ Claude
- Source
- youtube.com
- Author
- Anthropic
- Date
Context
A Google Cloud developer advocate lays out, in an Anthropic Code with Claude session, a four-part stack for building agents with Claude on Vertex AI: the open-source Development Kit for building and evaluating agents, compatibility with the for tools and , Vertex AI Agent Engine for deploying and managing them, and the Agent2Agent protocol for communication between agents built on different frameworks.
The speaker's case is that prototypes are easy but production is hard, because tooling is fragmented, agents from different frameworks do not talk to each other easily, and running them needs monitoring and logging. The demo agent is a birthday planner built on Claude 3.7 Sonnet that needs just three files. Claude 4, the speaker says, became available in Vertex AI's model garden that morning, with Agent Development Kit support promised in the coming weeks. This is a vendor walkthrough, not an .
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
Checking sign-in…
Loading comments…






