generated: '2026-07-18'
method: derived
source: openapi/cashflo-data-ingestion-openapi.json
api: CashFlo Data Ingestion API
authentication:
  style: JWT bearer token supplied in the Authorization request header
  scheme_name: jwt
  location: header
  parameter: Authorization
  see: authentication/cashflo-authentication.yml
idempotency:
  supported: true
  mechanism: payload-level deduplication
  field: skipDuplicates
  scope: per ingestion batch (purchase orders and GRNs)
  detail: >-
    Both ingestion endpoints accept a boolean `skipDuplicates` flag on the
    request body. When true, records already present are skipped rather than
    duplicated, making re-submission of the same batch safe. There is no
    Idempotency-Key header; deduplication is payload-driven.
pagination:
  supported: false
  detail: No list/collection endpoints; the batch-error endpoint returns the full error set for one batchId.
batching:
  supported: true
  detail: >-
    Ingestion is batch-oriented. A single request carries a `buyerOrgId` plus an
    array of records (`purchaseOrders[]` or `grns[]`). The API responds 202
    Accepted with a tracking `batchId`.
request_tracing:
  mechanism: batchId
  detail: >-
    Every accepted ingestion returns a `batchId` (IngestionSuccess.batchId) used
    to retrieve per-record failures via GET /v1/ingest/errors?batchId=...
versioning:
  scheme: uri-path
  current: v1
  detail: All endpoints are namespaced under /v1/ingest/.
error_handling:
  style: HTTP status codes (400/401/404/500); row-level errors via batch-error endpoint
  see: errors/cashflo-problem-types.yml
rate_limit_signaling:
  documented: false
tenancy:
  field: buyerOrgId
  detail: Records are scoped to a buyer organization via the required buyerOrgId field.