Introducing Real-Time People and Company Data

Two things are shipping today. First, real-time endpoints for people and companies that return a record fetched live at the moment you ask, not one served from cache. Second, a hard ceiling on how old any record in the lake is allowed to be: ninety days, enforced by default, for every profile and company we store.

The two are meant to be read together. Most of the time the lake is fast enough and fresh enough that you never think about where the answer came from. When a decision genuinely depends on the data being live, the real-time endpoint is there. When it does not, the ninety-day ceiling is the guarantee that the lake is never quietly serving you something stale.

The standard endpoints read from our own data, which is kept fresh by a continuous ingest and by a refresh that runs before answering any request for a record older than ninety days. That is the right default for almost every workflow, because it is fast and it does not depend on a third party being up.

Two endpoints bypass that path entirely: /person/realtime and /company/realtime. They query the live source directly at the moment you call, so the answer reflects what is true right now, not what was true when the record was last written. Use them when the difference between "current" and "a few days old" is the difference between a correct decision and a wrong one.

Both real-time endpoints are metered at $0.001 per call. There is no minimum commitment, no credit balance to buy in advance, and no tiering: one call, one tenth of a cent, invoiced against the same workspace that already holds your license.

Billing is deliberate about one edge case. A call that reaches the live source and gets an answer is billed, whether the subject was found or not, because the lookup was performed. A call that fails on our side, a timeout or an outage, is never billed. You only pay for what actually happened.

Every profile and company in the lake is refreshed within ninety days, by default, with no action required from you. If you request a record that is older than ninety days, it is refreshed before the answer is returned, so the number you act on is never older than the ceiling.

Ninety days is not an arbitrary round number. It is the cycle we already commit to in our terms, and it is short enough that the data you rely on for a pipeline or a report has a bounded age you can reason about, rather than an unknown one you have to check.

The real-time endpoint takes the same identifiers as the standard one, so it drops into an existing integration without a second data model:

The response is the same record shape as a standard profile lookup, so the rest of your pipeline does not change. The only difference is that the data behind it was fetched live, and that the call is metered.