Data API

How a data API works

Every API moves data, but a data API exists mainly to hand records to the caller. A payments API runs a charge. A messaging API sends a message. A data API answers questions such as who works at this company, what this domain runs or which posts mention this topic.

LexisNexis describes the interaction in four steps 1.

• Request. The client calls an endpoint, usually with GET or POST.
• Processing. The API authenticates the caller, checks permissions and retrieves the data.
• Response. The API returns normalized data, commonly as JSON or XML.
• Integration. The application feeds the data into dashboards, models or automated workflows.

The source behind the endpoint varies by provider. Some keep their own database and serve lookups from it. Others route each request to licensed third-party sources at request time and return whatever answers. RichAPI is in the second group. It states that contact and company lookups are fetched live from licensed third-party data sources, while web-scraping endpoints cache results server-side 3.

That design choice shows up in three places.

• Freshness. A live lookup reflects the source at call time, while a stored database reflects its last refresh, which is where data decay creeps in.
• Coverage. One database has one set of gaps, while waterfall enrichment across several sources can fill more of them.
• Cost. The billing rule decides whether an empty answer is free or billed.

Description files also matter. The OpenAPI Specification defines a standard, language-agnostic interface description for HTTP APIs, so people and computers can discover what a service does without reading its source code or inspecting traffic 2. The same document can feed documentation tools, code generators and test tools 2. RichAPI describes its API as plain REST with an OpenAPI spec and no SDK required 3.

Receipts are the last piece. A data API that reports what each call did makes spend auditable. RichAPI returns an execution log on every waterfall call that shows what answered, what was tried, how long each step took and whether the call was billed 3.

How to evaluate a data API before you commit

Five checks separate a data API that fits a pipeline from one that only demos well.

**Inputs.** Check which identifiers each endpoint accepts, such as a profile URL, a domain or a name plus company, and what happens when one is missing. A provider that only accepts one identifier forces extra lookups before the useful call.

**Freshness.** Ask whether records are fetched live or served from a stored copy, and how old a stored copy can get. Stale fields lower data accuracy even when match rate looks healthy.

**Billing rule.** Read whether each endpoint bills per call or per result, and what an empty answer costs. RichAPI publishes per-call, per-result and hybrid models, with status checks and inquiry endpoints free 4. Its email and phone lookups that find nothing cost nothing 4.

**Receipts.** Look for per-call evidence of what ran. RichAPI pairs its execution log with a free usage endpoint that aggregates spend per endpoint each month 3.

**Provenance.** Ask where the data comes from and who is responsible for collecting it. RichAPI names licensed third-party sources for contact and company lookups, public web sources for scraping endpoints and the customer's own inputs for AI enrichment 3.

A small test on a frozen sample answers most of these checks in an afternoon. Pick rows the team already knows well, run them through the candidate API and compare what comes back with what is true. RichAPI signup includes 100 free credits for that first test 3.

Example of a data API filling CRM account fields

A sales team wants account context in the CRM before the first call, so reps stop opening five tabs to research a company. The ops lead wires a data API into the new-account workflow.

Trigger         new account created in the CRM
Call 1          enrich_company, 1 credit per call
Call 2          web_tech_stack, 1 credit per call
Output          JSON fields mapped to account properties
Empty answer    still billed on these two endpoints
Spend check     usage endpoint, free

Each new account triggers two calls. Company enrichment and tech stack detection each cost 1 credit per call on the published rate card 4. At 400 new accounts a month, both lookups together cost at most 800 credits 4.

Both endpoints bill on a successful call, so a thin profile still costs its credit 4. The workflow filters out junk domains and test accounts before calling, which keeps spend tied to real prospects.

When a rep opens the account, firmographics such as headcount and industry and technographics such as the detected stack are already there as fields. Nobody imports a spreadsheet once a quarter, and no field is older than the account itself.

Once a month, the ops lead compares spend from the free usage endpoint with the number of accounts enriched 3. Cost per enriched account stays visible, and a sudden jump points to a loop or a bad trigger before it reaches the invoice.

Data API vs data feed

Both deliver third-party data into a company's systems. A data API answers on demand. A data feed typically delivers scheduled bulk updates 1.