Verify Workloads with the VB-OS Python SDK

verify

Submit a single workload for evaluation against a deployed boundary.

verify(
    workload: dict,
    project: str | None = None,
    boundary_ref: str | None = None,
    metadata: dict | None = None,
) -> EvaluationResult

Parameters

Parameter	Type	Required	Description
workload	dict	Yes	The evidence payload to evaluate
project	str | None	No	Project ID or slug
boundary_ref	str | None	No	Boundary reference to evaluate against
metadata	dict | None	No	Arbitrary metadata (non-authoritative, does not affect evaluation)

Returns

An EvaluationResult typed model containing the evaluation decision and details.

Example

from vbos import VBOSClient

with VBOSClient(api_key="vbos_your_key_here") as client:
    result = client.verify(
        workload={
            "npi_number": 1234567890,
            "license_state": "TX",
            "credential_count": 3,
        },
        project="healthcare-onboarding",
        boundary_ref="provider-credential-check",
    )

    print(result.decision)        # "ASSERT" or "DEFER"
    print(result.evaluation_id)

With metadata

result = client.verify(
    workload={"amount": 15000, "region_code": 4},
    project="loan-origination",
    metadata={"source": "web-portal", "operator_id": "op-42"},
)

Metadata is stored with the evaluation record but never influences the evaluation outcome.

verify_batch

Submit multiple workloads in a single request.

verify_batch(
    workloads: list[dict],
    project: str | None = None,
    boundary_ref: str | None = None,
) -> dict

Parameters

Parameter	Type	Required	Description
workloads	list[dict]	Yes	List of workload payloads
project	str | None	No	Project ID or slug
boundary_ref	str | None	No	Boundary reference

Example

results = client.verify_batch(
    workloads=[
        {"amount": 500, "risk_score": 12},
        {"amount": 75000, "risk_score": 88},
        {"amount": 1200, "risk_score": 45},
    ],
    project="lending-compliance",
)

for item in results["evaluations"]:
    print(item["evaluation_id"], item["decision"])

validate

Validate a workload against a boundary without recording an evaluation. Useful for testing and development.

validate(
    workload: dict,
    project: str | None = None,
    boundary_ref: str | None = None,
) -> dict

Parameters

Parameter	Type	Required	Description
workload	dict	Yes	The evidence payload to validate
project	str | None	No	Project ID or slug
boundary_ref	str | None	No	Boundary reference

Example

validation = client.validate(
    workload={"amount": 500, "risk_score": 12},
    project="lending-compliance",
    boundary_ref="loan-limits",
)

print(validation["decision"])   # "ASSERT" or "DEFER"

The validate method runs the same evaluation logic as verify but does not create an evaluation record. Use it to test boundary behavior during development or to preview results before committing.