Shared keys and copied tokens
Agents run on shared API keys or human tokens, so no one can say who did this, or for whom.
For AI & platform teams
Put agents into production without borrowed identities.
Your first agents and MCP tool servers are going live. HelixIAM gives each one an owner, a lifecycle and delegated authority that only narrows — on your Kubernetes, next to the IdP you already have.
Request · agent acting for alice
POST /realms/acme/oauth2/token grant_type=urn:ietf:params:oauth:grant-type:token-exchange subject_token=eyJ… # alice actor_token=eyJ… # agent: invoice-assistant scope=invoices:read
Issued access token · decoded
{
"sub": "alice",
"act": { "sub": "agent:invoice-assistant" },
"scope": "invoices:read",
"aud": "https://api.acme.example/invoices",
"iss": "https://id.acme.example/realms/acme"
}actThe agent is named in every token; the user stays the subject.scopeAuthority narrows to what the agent needs. It never widens.ownerSuspend the agent and it gets no new tokens.
What blocks agent projects today
No way to stop just one
Stopping one agent means rotating a key that other systems depend on.
Secrets in pipelines and pods
Long-lived credentials in CI and Kubernetes become an audit finding waiting to happen.
What you get
Accountable agents
Each agent has an owner and a lifecycle; agents whose owner has left are flagged.
- Delegated tokens record user and agent
- Authority only narrows
- Suspend an agent and it gets no new tokens
Keyless workloads
Kubernetes and CI workloads exchange their platform token for a short-lived HelixIAM token.
- No long-lived secrets
- Bound to application roles
- Sender-constrained tokens with DPoP
An audit trail per agent
Every agent token issued or refused, direct or delegated, plus lifecycle changes and admin actions, in a searchable audit log; workload token issuance is audited too.
- Stream to your SIEM over HTTPS
- HMAC-signed webhooks
- Ready for MCP (RFC 9728)
Questions
Do we have to replace our identity provider?
No. People keep signing in through your existing IdP and are brokered into HelixIAM over OpenID Connect or SAML; HelixIAM governs agents and workloads.
How does a pilot work?
A paid pilot runs on your own Kubernetes, with success criteria agreed up front and a production plan at the end.
Early access/now onboarding launch partners
See it on a running system.
Start with a technical demo and a paid pilot on your own use case, and work directly with the engineers who build HelixIAM.