Vertex AI Agent Builder

Connect any agent, anywhere with the open Agent2Agent (A2A) protocol. This universal communication standard enables agents across different ecosystems to communicate with each other, irrespective of the framework (ADK, LangGraph, Crew.ai, or others) or vendor they are built on. Using A2A, agents can publish their capabilities and negotiate how they will interact with users (through text, forms, or bidirectional audio/video)–all while working securely together.

Like APIs but for agent communication, A2A lets you turn isolated agents into collaborative teams, freeing you from concerns about investing in incompatible agent frameworks. A2A is a collaborative effort to standardize agent interoperability with contributions from a growing ecosystem of 50+ partners including Box, Deloitte, Elastic, Salesforce, ServiceNow, UiPath, UKG, and many more.

You can equip agents with your existing enterprise truth data without building from scratch, using any approach you prefer: 

ADK supports Model Context Protocol (MCP), so your agents connect to the vast and diverse data sources or capabilities you already rely on by leveraging the growing ecosystem of MCP-compatible tools.

You can also connect your agents to enterprise systems through 100+ pre-built connectors, your custom APIs in Apigee, or workflows in Application Integration. These easy-to-use building blocks ensure your agents behave in alignment with your business rules and operate within appropriate guardrails.

Using ADK, your agents can also tap into your existing API investments managed in Apigee API Management–irrespective of where they reside.

Simplify your path to production with Agent Engine, our fully managed runtime that eliminates deployment complexity. Agent Engine handles infrastructure management, scaling, security, and monitoring so you can focus on agent capabilities rather than operational concerns.

You can deploy agents seamlessly regardless of which framework or model provider you've chosen. With Agent Engine, you can also keep the context in your conversations so your interactions with agents feel human-like, rather than starting from a blank slate each time.

Agent Engine supports short-term memory and long-term memory. This way, you can manage your sessions and your agents can recall your past conversations and preferences.

Lastly, with comprehensive evaluation tools and Example Store from Vertex AI you can improve agent performance and refine based on real-world usage.

Equip your agents with comprehensive retrieval-augmented generation (RAG) capabilities that let them intelligently access and interpret your organization's knowledge. Vertex AI Search provides an out-of-the-box RAG solution that gets you started with just a few clicks, while Vector Search supports hybrid techniques that combine vector-based and keyword-based approaches for more relevant responses. You can also build your own custom RAG engine by connecting to diverse data sources including local files, Cloud Storage, Google Drive, Slack, Jira, and many more.