- CollectionAgent Runtime & Infrastructure
- Classificationagent-native
- MCPUnavailable
- Agent SkillUnavailable
- InterfacesREST · GRPC · SDK · A2A · http
- Verified2026-10-19
What it does
Gemini Enterprise Agent Platform (formerly Vertex AI Agent Engine) is Google Cloud’s managed layer for running agent workloads in production. Live docs: “a fully managed environment for developers to handle testing, release management, and reliability at a global scale” — Agent Runtime, Sessions, Memory Bank, Code Execution, Computer Use, plus evaluation and example-store quality loops. Framework guides on the scale page still cover ADK, LangChain, LangGraph, LlamaIndex, AG2, and Agent2Agent (A2A).
Agent-facing interfaces
- Catalog protocols: REST, GRPC, SDK, A2A, http
- MCP: ⚠️ No standalone MCP server
- Agent Skill: ⚠️ Not yet published
- Onboarding type:
sdk
How agents onboard
pip install "google-cloud-aiplatform[agent_engines,adk]"
See the full onboarding section in the dossier.
Protocol surface
| Interface | Detail |
|---|---|
| Vertex AI SDK (Python) | google-cloud-aiplatform with agent_engines — primary path in docs |
| REST / gRPC | Vertex AI platform APIs for ReasoningEngine / agent resources |
| OpenTelemetry | Tracing integration per runtime tracing docs |
| A2A | Agent2Agent protocol guides remain linked from the scale hub |
Verification
Last verified 2026-10-19.
Sources:
- https://github.com/googleapis/python-aiplatform