Random records API
Sample synthetic records conforming to the specified record schema.
CLI equivalent: simhealth sample {recordType} --count 1 --seed 1842 (--format json|ndjson|avro, extra --constraint key=value).
The project ID must match the project permanently bound to the Bearer API key. The unscoped path GET /v1/random/{recordType} is retained as a legacy compatibility route for existing clients; use the project-scoped path for new integrations.
Request parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
recordType | path string | required | Record definition name, such as person or encounter. |
count | query integer | 1 | Number of records. Up to 10,000 for JSON and 10,000,000 for Avro. |
seed | query integer | random | 64-bit unsigned seed for deterministic sampling. |
fields | query string | all | Comma-separated fields to include. |
Headers and content negotiation
Authorization: Bearer <SIMHEALTH_API_KEY>is required. A signed-in Firebase ID token or an API key is enough. Omittingx-customer-iddoes not return 400.x-customer-idis an optional credit-ledger override. When omitted, the call is metered asanonymous. Ansk_live_...value in this header is ignored for the ledger. Any other value must be 1–128 ASCII letters, digits,_, or-, must not beanonymous, and must not start withcs_. An invalid value is HTTP 400{"error":"invalid customer id","message":"invalid customer id"}.Accept: application/jsonreturns buffered JSON (the default), up to 10,000 records.Accept: application/avro-binaryorapplication/avroreturns a streaming Avro OCF, up to the caller's record quota (default ceiling 10,000,000).Accept: application/x-ndjsonreturns newline-delimited JSON.
The API does not guarantee a throughput rate or a client rate limit. The figures above are count caps.
Example response
{
"recordType": "person",
"count": 1,
"seed": 1842,
"records": [
{
"person_id": "P-981742",
"first_name": "Maria",
"last_name": "Santos",
"gender": "female",
"birth_date": "1984-06-12",
"mrn": "MRN-004928"
}
]
}