Announcing Dome Managed Agents.

Try it now

Google sign-in and Dome

People sign in with Google, and Dome checks that identity on every call. An agent sends the user's Google ID token with each call. Dome verifies it against Google, and rules decide on that person's email and Workspace domain. Audit records the agent and the person, every time.

Dome provides verified identityand access control with Google.

Verified against Google

Dome checks the ID token's signature against Google's keys, its issuer and its expiry. Pin the audience to your OAuth client ID.

Rules read email and domain

The person's email reaches every rule. For Workspace accounts, so does the hd claim with their domain.

No groups in the token

Google ID tokens carry no groups. Scope rules by email, domain or subject instead.

Googlein three steps.

Register Google as a verification provider, have the agent pass the user's ID token, and scope rules by domain. Dome needs no Google credentials of its own.
  1. Register Google as a verification provider

    The issuer is the same for every Google account. Pin the audience to your app's OAuth client ID.

    shell
    dome verification-providers create \  --name google \  --method oidc \  --oidc-url https://accounts.google.com \  --oidc-expected-audience <client-id>.apps.googleusercontent.com
  2. Pass the user's ID token with each call

    Send the ID token from sign-in, with the email scope. Google access tokens aren't JWTs, so Dome can't verify them.

    python
    from openai import OpenAI
    client = OpenAI(    base_url=f"{GATEWAY_URL}/v1",    api_key=DOME_AGENT_KEY,    default_headers={"X-Dome-Act-As": google_id_token},)
  3. Scope by Workspace domain

    Google sets hd only for Workspace and Cloud organization accounts, and it arrives under principal.act_as.claims.

    cedar
    principal has act_as &&principal.act_as has claims &&principal.act_as.claims has hd &&principal.act_as.claims.hd == "example.com"

A ruleon the person's email.

Only people with an example.com address may use the agent's GitHub tools. A personal Gmail account is refused, even with a valid token.
cedar
forbid (principal, action == Dome::Action::"mcp:call", resource is Dome::MCPTool)when { resource.connection_name == "github" }unless {  principal has act_as &&  principal.act_as has email &&  principal.act_as.email like "*@example.com"};

One call,two outcomes.

Switch the caller or the argument and watch the same call decide differently, with every decision landing in audit.

Acting for

agent code-reviewer · acting as a.okafor@example.com
github/search_code(query: "rate limit repo:acme/web")
  1. CallerID token verified against Google
  2. Emaila.okafor@example.com
  3. RuleGitHub tools are open to example.com
DecisionAllowed

Bringing it alltogether.

Connecting Google to registered agents, tools, and models in Dome completes a governed agent application.
Dome

Acting for

GoogleIdentity

Control point

GatewayEvery call decided and audited
RulesGuardsQuotas

Common questionsabout Google.

Do Google ID tokens include groups?
No. Rules can use the person's email, subject and, for Workspace accounts, the hd domain claim.
Can an agent send a Google access token?
No. Google access tokens aren't JWTs, so the agent sends the ID token instead.
Will any Google account get through?
Any account your OAuth app lets sign in gets a token that verifies. Rules on email or hd decide which accounts may call anything.

More of whatDome works with.

Model

Claude Fable

Fable 5.1 from Anthropic and Amazon Bedrock in one failover pool, open to one group and capped by quota.
Provider

Anthropic

The Claude API behind the Model Broker: the key held in Dome, every call authorized, metered and audited.
MCP server

Atlassian

The Atlassian Rovo MCP server behind the Tool Gateway, with rules that decide each Jira and Confluence call on its tool and site.
Runtime

LangGraph

LangGraph agents with their model calls on the Model Broker and their MCP tools on the Tool Gateway.
Client

Claude Code

Claude Code on a Dome Gateway with per-developer sign-in, rules on every tool call, and audit by name.
Agent service

TinyFish

TinyFish's web agents behind the Tool Gateway, with rules that decide each run on the site it targets.

Talk with ourFDE team.

Or start free, and register your first agent in minutes. Connect it through a gateway, give it rules, and see every call it makes in one audit trail.

Talk with our FDE team

Our engineers design, build and run governed agents for the part of your business you want to change. You see what they deliver.

Identity providers for AI agents: a verified person on every call

Your identity provider already knows your people. Dome verifies that identity on every agent call, so rules and audit name the person an agent acted for.

Every agent you run, under one system of control.