# Global Data Check

Global Data Check matches a person against the **Global Data Universe** and returns a per-field breakdown of how well the supplied details line up with one or more candidate records.

It is a single endpoint - `POST /globaldata_check` - that supports a wide range of identity verification, contact validation, and lookup workflows. The formal request and response schema (with every field, default, and enum) lives in the [Global Data Check API reference](/docs/reference#global-data-check); this guide focuses on **how to use it** and walks through worked sandbox examples for each common use case.

The check runs against the Global Data Universe, an in-house dataset of person and contact records aggregated from many sources.

The response is **match information only - no PII is returned.** Every field in `match_results` is a status flag (`match`, `no_match`, `match_year`, etc.) describing how the supplied input compares to a candidate. The candidate's underlying details (their actual DOB, phone, address, etc.) are never sent back. This makes the check a clean choice for confirmation flows where you already hold the data you are checking and just need a yes/no/partial signal.

The check is also **very fast** and is suitable for high-volume bulk processing - for example, scoring an entire customer database overnight or running real-time checks at signup throughput:

| Percentile | Response time |
|:-----------|--------------:|
| p50        | 180 ms        |
| p75        | 550 ms        |

Per-request times vary with how much narrowing the supplied identifiers provide - a request with a phone, email, DOB, or full address resolves on the fast path; broad searches against very common last names with no other discriminator make up the slower tail.

Common things you can do with it:

- Confirm a person exists with a given name and date of birth (KYC / identity verification).
- Confirm that a phone number or email belongs to a particular person (signup flows, fraud checks).
- Verify someone lives at a given address, with full or partial address inputs.
- Find candidate matches for a partially-known identity to support manual review queues.
- Match against typos, nicknames, and reversed dates of birth without a separate fuzzy-matching pipeline.
- Score the match confidence and use it to drive auto-approve / manual review / reject routing.