Agent Registry lets you register AI agents so that developers and orchestrator agents can discover and invoke their capabilities across your AI ecosystem. By publishing your agent's A2A skills, you help other developers and orchestrator agents in your organization find and reuse your work.
Agent Registry integrates with versions 0.3 and 1.0 of the open source
Agent2Agent (A2A) protocol, which lets
agents declare their capabilities and identity in an
Agent Card. Depending on where your
agents run, you can register them using
automatic registration or
manual registration.
Automatic registration
Supported Google Cloud runtimes register your agents automatically, and built-in Google agents are discoverable without configuration. Within the governance project where Agent Registry is enabled, automatic registration happens either by default or through an opt-in deployment step. The registration process operates at the project level and doesn't automatically discover remote agents deployed in other Google Cloud projects.
For information about Google Cloud runtimes that support automatic registration and the metadata indexed for automatically registered agents, see Use automatic registration for agents.
Manual registration
Agent Registry uses a central Google Cloud governance project. The term remote refers to agents hosted outside the governance project in separate Google Cloud projects, on-premises environments, or external platforms, or running on Google Cloud runtimes that don't support automatic discovery. You must use manual registration to add remote agents to the governance project's registry.
To manually register an agent, use the Google Cloud console, the Google Cloud CLI,
Terraform, or the Agent Registry API to create a Service resource that stores the
agent's metadata, including its endpoint URL and configuration. After you create
the Service resource, Agent Registry automatically creates a read-only
Agent resource that you can use for
discovery.
To learn how to manually register A2A-compliant and standard REST agents, see Use manual registration for agents.
Cross-project registration considerations
If your organization uses a centralized governance architecture with Agent Gateway, such as the multi-project pattern described in Sample deployment patterns, you can deploy a gateway in Agent-to-Anywhere egress mode in a central governance project where Agent Registry is enabled, while deploying agents in separate Google Cloud projects.
Because automatic registration only discovers resources in the governance project where Agent Registry is enabled, it doesn't automatically discover remote agents deployed in other Google Cloud projects. To let the central gateway locate and govern those remote agents, you must manually register each remote agent in the central governance project's registry. For instructions, see Register an agent from another project.
Integrate standalone skills
Agent Registry lets you manage and govern standalone
skills for agents as top-level resources.
Standalone skills are registered as Skill resources in your registry,
letting administrators apply fine-grained access policies and manage version
history using
skill revisions.
When building and managing agents in Gemini Enterprise, you can reference these registered standalone skills to extend your Gemini Enterprise agent's capabilities. For information on how to register and manage standalone skills, see Register skills and Manage skills.