Vertex AI Agent Engine
App : A product you use directly
Google Cloud managed runtime for deploying agents: execution, sessions, long-term memory (Memory Bank), code execution, and agent identity.
Google's counterpart to AgentCore, designed for agents written with ADK. Memory Bank isolates memories per user, with configurable lifetimes and IAM permissions.
Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.
What it does, concretely
Deploying an AI agent in production means managing infrastructure, scaling, context persistence, and execution security. Vertex AI Agent Engine, now part of the Gemini Enterprise Agent Platform, handles these concerns within a fully managed Google Cloud environment.
In practice, the service offers four main building blocks: a serverless runtime to deploy and scale agents without managing servers, a sessions system to maintain conversation state, a Memory Bank to store persistent information across multiple sessions, and an isolated Python code execution environment. Agents can be written with ADK, LangChain, LangGraph, LlamaIndex, AG2, or the Agent2Agent protocol.
The platform targets development teams building enterprise-grade AI agents who need observability, IAM access control, data residency, and enterprise security features such as VPC Service Controls or customer-managed encryption keys (CMEK).
It is not suited for those looking for a no-code solution or a consumer-facing tool: using it requires familiarity with Google Cloud, configuring service accounts, and writing code to deploy agents.
What it offers
- ·Managed serverless runtime to deploy and scale agents without managing infrastructure, with sub-second cold starts (paid).
- ·Sessions to maintain conversation state and context within a single user-agent interaction (paid).
- ·Memory Bank to persistently store user preferences and facts across multiple sessions, with access control via IAM Conditions (paid).
- ·Python code execution in a secure, isolated sandbox environment for computation and automation (paid).
- ·IAM agent identity to assign precise permissions to each deployed agent.
- ·Compatibility with multiple frameworks: ADK, LangChain, LangGraph, LlamaIndex, AG2, and Agent2Agent.
- ·Built-in observability through Cloud Trace, Cloud Logging, and custom metrics to monitor agents in production.
- ·Enterprise security features: VPC Service Controls, CMEK, data residency, and HIPAA compliance for most services.
In practice
- Price
- Paid
- Commercial use
- Commercial use: yes
- Available on
- cloud
- FR interface
- No
Security and data
- Rating
- Security: one caveat, see the card
- Your data
- Data with the vendor
- Maintenance
- Active
Links and repository checked on 25 October 2026 · added on 27 September 2026
How to install Vertex AI Agent Engine
pip install google-cloud-aiplatform[agent_engines,adk] puis déploiement depuis le SDK (projet Google Cloud requis)