Skip to content
Halo by OmniTrust

Model providers

Available now

Halo and OpenAI

Govern OpenAI model access through Halo’s model gateway, with scoped credentials and policy controls.

Book a walkthrough Start with Halo

How Halo works with OpenAI

Halo routes approved OpenAI model requests through the model gateway, where identity, policy, budgets, and evidence controls apply to agent traffic.

Teams can use OpenAI through bring your own key, with credentials bound at the gateway. The configured model and policy determine what a workflow may do.

Governance

What Halo governs for OpenAI

Model access

Route approved requests through Halo’s model gateway under the selected access mode.

Policy checks

Evaluate model requests and responses in supported gateway paths against configured policy.

Brokered credentials

Keep provider credentials at the gateway boundary and give agents scoped capabilities instead of raw keys.

Budget and evidence

Apply account or agent budgets and retain reviewable decision and cost evidence.

Example agent jobs

Put agents to work with OpenAI

Illustrative jobs. Each one runs with scoped access, policy, and evidence.

Illustrative support response

For example, a support agent could draft a response using an approved OpenAI model, with policy checks and an approval gate for higher-impact actions.

Illustrative internal summary

For example, an internal agent could summarize a request using an approved model, subject to configured policy and budget controls.

Setup

Get connected in three steps

  1. 1

    Choose the provider

    Select OpenAI when you choose bring your own key for an agent.

  2. 2

    Bind the credential

    Add the provider key through Halo’s credential flow so the gateway can broker it for the agent.

  3. 3

    Set controls

    Choose an approved model, configure policy and budgets, and test the agent workflow.

FAQ

Questions, answered

Yes. OpenAI is a supported BYOK provider, and Halo brokers the credential through its gateway.

Model requests and responses can be checked in supported gateway paths, with decisions and costs available as reviewable evidence.

Related capabilities

Model GatewayManaged and BYOK model accessLearn more about Model GatewayModel RoutingRoute requests to the right approved modelLearn more about Model RoutingRuntime GuardrailsDetect sensitive data and injectionLearn more about Runtime Guardrails

More model providers

Put a halo around the AI your business runs.

See discovery, runtime policy, approvals, and evidence working together on your own use cases.

Book a walkthrough Start with Halo