How it works
When you call a pass-through API, the provider authenticates with the source platform, translates your request, fetches or writes the data live and returns the response. Nothing is persisted beyond logs and metadata. This keeps customer data out of an extra system and always returns current values, but each read is limited by the source's speed, rate limits and availability.
Example
An underwriting tool fetches a borrower's balance sheet at the moment of application. Pass-through mode returns the ledger's current figures and stores nothing, which suits the lender's data-minimization policy.
In integrations
Pass-through and sync-and-store each fit different objects. Unify Int is designed to let you choose per object: pass-through reads for sensitive or fast-changing data, stored sync for large histories and fast queries. This is different from a passthrough request, which reaches a native endpoint the unified model does not cover. Choose pass-through when freshness or data minimization matters more than query speed, for example a one-time underwriting pull. Expect slower bulk reads, and plan for source outages because there is no stored copy to fall back on.