Agent Development Kit (ADK) Integrations

Gemini Enterprise Agent Platform is a comprehensive solution from Google Cloud designed to help organizations build, scale, govern, and optimize AI agents. It represents the evolution of Vertex AI, combining advanced model development with new capabilities for agent orchestration and integration. The platform provides access to over 200 leading AI models, including Google’s Gemini series and third-party options like Anthropic’s Claude. It enables teams to create intelligent agents using both low-code and code-first development environments. With features like Agent Runtime and Memory Bank, businesses can deploy long-running agents that retain context and perform complex workflows. The platform emphasizes security and governance through tools like Agent Identity, Agent Registry, and Agent Gateway. It also includes optimization tools such as simulation, evaluation, and observability to ensure consistent agent performance.

Cloud Run is a fully-managed compute platform that lets you run your code in a container directly on top of Google's scalable infrastructure. We’ve intentionally designed Cloud Run to make developers more productive - you get to focus on writing your code, using your favorite language, and Cloud Run takes care of operating your service. Fully managed compute platform for deploying and scaling containerized applications quickly and securely. Write code your way using your favorite languages (Go, Python, Java, Ruby, Node.js, and more). Abstract away all infrastructure management for a simple developer experience. Build applications in your favorite language, with your favorite dependencies and tools, and deploy them in seconds. Cloud Run abstracts away all infrastructure management by automatically scaling up and down from zero almost instantaneously—depending on traffic. Cloud Run only charges you for the exact resources you use. Cloud Run makes app development & deployment simpler.

Gemini is Google’s advanced AI assistant designed to help users think, create, learn, and complete tasks with a new level of intelligence. Powered by Google’s most capable models, including Gemini 3, it enables users to ask complex questions, generate content, analyze information, and explore ideas through natural conversation. Gemini can create images, videos, summaries, study plans, and first drafts while also providing feedback on uploaded files and written work. The platform is grounded in Google Search, allowing it to deliver accurate, up-to-date information and support deep follow-up questions. Gemini connects seamlessly with Google apps like Gmail, Docs, Calendar, Maps, YouTube, and Photos to help users complete tasks without switching tools. Features such as Gemini Live, Deep Research, and Gems enhance brainstorming, research, and personalized workflows. Available through flexible free and paid plans, Gemini supports everyday users, students, and professionals across devices.

Agent2Agent (A2A) is a protocol developed by Google to enable seamless communication between AI agents. It facilitates the transfer of knowledge and tasks between different AI systems, allowing them to collaborate and execute complex workflows. A2A aims to enhance interoperability between AI agents, enabling more sophisticated, multi-agent systems that can perform tasks autonomously across various platforms and services.

Gemini Enterprise app is an advanced AI-powered platform that brings Google’s AI capabilities to every employee, enabling organizations to automate workflows, analyze data, and create high-quality content across multiple business functions. It securely connects to tools like Microsoft 365, Google Workspace, HubSpot, and Jira, allowing users to search and interact with their business data using natural language. The platform supports prebuilt agents such as NotebookLM and Deep Research, helping teams quickly extract insights and streamline tasks. It also allows users to build custom no-code agents to automate multi-step workflows across different applications. With centralized management, organizations can deploy and monitor all agents from a single interface. Built-in security and governance features ensure data privacy and compliance with enterprise standards. Overall, Gemini Enterprise app enhances productivity by combining AI automation with secure data integration.

asqav is an AI governance and security platform designed to make AI agents audit-ready by providing real-time monitoring, enforcement, and verifiable proof of every action taken by an agent. It introduces a lightweight SDK that allows developers to integrate governance directly into their agents in just a few lines of code, enabling continuous oversight across the full lifecycle of AI operations. It includes behavioral monitoring to detect issues such as drift, rate limits, and scope violations, along with advanced threat detection that identifies prompt injections, exposure of sensitive data, toxic outputs, and other risks. It enforces policy through configurable “policy gates,” which apply per-agent rules, preflight checks, and dynamic approvals before actions are executed, ensuring that agents operate within defined boundaries. asqav also provides automated incident response capabilities, including the ability to suspend, quarantine, or escalate risky agents.

Agentspan is an open source server and SDK designed to bring durable execution to AI agents, transforming how they run in real-world environments beyond simple demos. It allows developers to define agents in Python and compile them into persistent, crash-safe workflows where execution state lives on the server rather than in the local process, ensuring that work is never lost if a system crashes or restarts. This architecture enables agents to pause, resume, and continue from the exact step they left off, even when reconnected from a different machine. It supports human-in-the-loop workflows, allowing agents to halt for approval and resume seamlessly through interfaces like Slack, web portals, or code. It also enables multi-agent pipelines, where several agents can be chained together in a single expression, with each step logged, observable, and recoverable across the entire workflow.

AG-UI is an open, lightweight, event-based protocol that standardizes how AI agents connect to user-facing applications. Built for simplicity and flexibility, it enables seamless integration between AI agents, real-time user context, and user interfaces. AG-UI is designed for agent-human interaction: during agent executions, backends emit events compatible with standard AG-UI event types, and agent backends can accept simple AG-UI-compatible inputs as arguments. It works with any event transport, including SSE, WebSockets, webhooks, and other streaming systems, while providing a flexible middleware layer that ensures compatibility across diverse environments. AG-UI brings agents into user-facing applications and complements the wider agentic protocol stack: MCP gives agents tools, A2A allows agents to communicate with other agents, and AG-UI connects agents directly to the user interface.

Agent Control is the open source control plane for AI agents, built to establish a new standard for governing agent behavior at scale. It solves the problem of scattered, hardcoded checks by giving teams a centralized governance layer with step-level enforcement that can be managed from a single control plane and updated in real time without touching agent code. Developers can make any function governable by adding the control() decorator, turning meaningful decision points inside an agent into independently governed control points with their own policies. When a decorated function executes, Agent Control evaluates the input or output against the active policy and returns a decision: deny, steer, warn, log, or allow. If the decision is denied, the SDK raises a ControlViolationError before the unsafe action can proceed. Policies are decoupled from code, so developers decide where to place control hooks while policy teams decide what those hooks enforce.

Silmaril is a self-healing prompt injection defense designed to protect AI systems from increasingly complex, multi-step attacks that traditional guardrails fail to stop. It operates by wrapping inference calls and evaluating whether an execution sequence is leading toward a harmful outcome, rather than simply filtering inputs. It uses a multihead classifier that analyzes user intent, application context, and execution states together, enabling it to detect indirect injection, multi-turn attack chains, context poisoning, and tool abuse before damage occurs. Silmaril continuously strengthens its defenses through autonomous threat hunting agents that probe systems, discover vulnerabilities, and generate synthetic training data from real attack scenarios. These insights are used to retrain the model automatically, deploying updated protections in under an hour and propagating anonymized defenses across all deployments.