FeaturesHow It WorksPricingDocs

Context Overrides & wrap()

Override budget guards, downgrade maps, and workflow metadata per call using contextvars (Python) or AsyncLocalStorage (Node).

Agent hint

Use for multi-tenant wrap(), per-request budget_guard overrides, and avoiding global config mutation.

In scope

  • wrap() / awrap()
  • Per-request budget_guard
  • downgrade_map overrides

Out of scope

  • patch_all double-counting — see Harness Matrix
  • Resilience retry — see Reliability
Python
res = auditor.wrap(
    call_fn=openai.chat.completions.create,
    build_params=lambda: {
        "model": "gpt-4o",
        "messages": [{"role": "user", "content": "Process request"}],
    },
    options={
        "workflow_id": "tenant_abc",
        "budget_guard": {"max_cost_per_session": 0.05, "on_budget_exceeded": "throw"},
        "downgrade_map": {"gpt-4o": "gpt-4o-mini"},
    },
)
Node.js
const response = await auditor.wrap(
  (params) => openai.chat.completions.create(params),
  () => ({ model: "gpt-4o", messages: [{ role: "user", content: "Process request" }] }),
  {
    workflowId: "tenant_abc",
    budgetGuard: { maxCostPerSession: 0.05, onBudgetExceeded: "throw" },
    downgradeMap: { "gpt-4o": "gpt-4o-mini" },
  },
);

See also