Gemini Enterprise Agent Platform (original) (raw)

Gemini Enterprise Agent Platform

Build, scale, govern, and optimize with Gemini Enterprise Agent Platform

Gemini Enterprise Agent Platform is our open and comprehensive platform that empowers businesses to rapidly build, scale, govern and optimize enterprise-grade agents grounded in your enterprise data. It provides the full-stack foundation and extensive developer choice you need to transform your applications and workflows into powerful, reliable agentic systems at global scale.

API quickstart Get started in Agent Studio Get an API key

Everything you need to buildproduction-ready agents

Build

• Build with Agent Platform
  Learn about the different tools and services you can use to build agents with Agent Platform.
  
• Agent Development Kit
  The Agent Development Kit (ADK) is a modular, model-agnostic framework designed to make building and deploying complex AI agents feel just like standard software development.
  
• Agent Garden
  Explore Google's library of pre-built agents.
  
• Agent Studio
  Low-code canvas for designing complex multi-agent reasoning loops.
  
• Skill Registry
  Skill Registry serves as a secure, private, and low-latency repository for managing and discovering agent skills.
  
• RAG Engine
  RAG Engine helps you securely connect your private data to LLMs, reducing hallucinations and improving answer accuracy by providing the model with relevant, context-rich information.
  
• Vector Search
  Vector Search is an AI-native search engine that combines storage and retrieval into one scalable, auto-tuned platform, simplifying how you store, search, and manage your data for AI applications.
  
• Managed Agents API
  A config-driven, REST-first API for building autonomous agents inside a fully managed sandbox.
  

Scale

• Deploy agents
  Deploy agents on Agent Engine using Python SDK, source files, or container images.
  
• Manage deployed agents
  Manage and update deployed agents, including scaling and resource allocation.
  
• Use an agent
  Interact with deployed agents using the API or SDK.
  
• Agent Platform Sessions
  Manage conversational sessions and state for multi-turn interactions.
  
• Agent Platform Memory Bank
  Store and retrieve long-term memories for agents to maintain context.
  
• Set up Memory Bank
  Configure and initialize the Memory Bank for your agents.
  
• Generate memories
  Extract and save memories from agent interactions.
  
• Fetch memories
  Retrieve stored memories to provide context to agent executions.
  

Govern

• Agent Gateway
  Centralized gateway for routing, securing, and monitoring agent traffic.
  
• Semantic governance
  Configure policies to ensure safe and compliant agent behavior.
  
• Content security
  Monitor and filter agent inputs and outputs for safety and security.
  
• Identity & IAM
  Manage access control and permissions for agents using Google Cloud IAM.
  
• Share agent
  Securely share agents across projects or organizations.
  

Optimize

• Agent evaluation
  Learn about evaluation concepts and how to assess agent performance.
  
• Observability overview
  Monitor agent health, latency, and resource usage.
  
• Traces
  Debug complex reasoning loops and tool calls with detailed execution traces.
  
• Topology
  Visualize agent architectures and dependency graphs.
  
• Offline evaluations
  Run evaluations on static datasets to measure quality and latency.
  
• Simulated evaluation
  Simulate user interactions and evaluate agent behavior against custom criteria.
  
• Online monitoring
  Continuously monitor and evaluate agent performance in production environments.
  
• Optimize prompts
  Auto-tune and optimize prompts to improve agent response quality and efficiency.
  

Built for developers,designed for enterprise

• Gemini
  Advanced reasoning-first models optimized for complex agentic workflows and coding. Features Gemini 3.1 Pro and Gemini 3 Flash.
  
• Veo
  High-quality video generation models that create cinematic video from text or image prompts. Features Veo 3.1.
  
• Lyria
  Generate full-length music tracks and audio clips from simple text and image prompts. Features Lyria 3.
  
• Partner models
  Access curated third-party models offered as managed APIs. Includes Claude, Grok, and Mistral.
  
• Open models
  Choose from a variety of open weight models and select the best serving option. Includes Llama, DeepSeek, Mistral, and Qwen.
  

Get started with Agent Platform

Start using Agent Platform to build, scale, govern, and optimize enterprise-grade agents today.

See how it works

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2026-06-15 UTC.