Okta and Dome
Know who every agent is acting for. An agent sends the user's Okta token with each call. Dome verifies it against Okta, and your rules decide on that person's groups. Audit records the agent and the person, every time.
Dome provides verified identityand access control with Okta.
Verified on every call
Dome checks the token's signature against Okta's keys, its issuer and its expiry. A forged or stale token is refused.
Rules read their groups
The user's Okta groups, roles and email reach every rule. Give research the frontier models and support the rest.
Required where it matters
Require a verified user on a workspace or a single agent. Allow only certain groups, emails or subjects.
Oktain three steps.
Register your Okta authorization server with Dome, put groups in the token, and have the agent pass the token on. Dome needs no Okta credentials of its own.
Register Okta as a verification provider
Dome reads Okta's signing keys from its discovery document. Pin the audience to your app's client id.
shelldome verification-providers create \ --name corporate-okta \ --method oidc \ --oidc-url https://<org>.okta.com/oauth2/default \ --oidc-expected-audience <client-id>Add a groups claim in Okta
Dome reads sub, email, roles and groups from the token's top level. Add a groups claim to the authorization server your app uses.
Pass the user's token with each call
The agent keeps its own Dome key and adds the person's Okta token as a header.
pythonfrom openai import OpenAI client = OpenAI( base_url=f"{GATEWAY_URL}/v1", api_key=DOME_AGENT_KEY, default_headers={"X-Dome-Act-As": user_okta_token},)
A ruleon Okta groups.
Frontier models are refused unless the person the agent acts for is in ml-research. The group comes straight from their Okta token.
cedar
forbid (principal, action == Dome::Action::"llm:invoke", resource is Dome::LLMModel)when { resource has tier && resource.tier == "frontier" }unless { principal has act_as && principal.act_as.groups.contains("ml-research")};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 research-agent · acting as a.okafor
llm:invoke(model: "claude-opus-5-5")
- CallerToken verified against corporate-okta
- Groupsml-research, engineering
- RuleFrontier models are open to ml-research
DecisionAllowed
Bringing it alltogether.
Connecting Okta to registered agents, tools, and models in Dome completes a governed agent application.
Acting for
OktaIdentity
Agent
Control point
GatewayEvery call decided and audited
RulesGuardsQuotas
Tools
Models
Common questionsabout Okta.
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
OpenAI Agents SDK
OpenAI Agents SDK agents with their models 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.