APIs that do
one thing well.
Email verification, on-page SEO audits, PDF extraction, text understanding and PII redaction. Each one is a single Worker with a clear contract, an OpenAPI document you can import, and a limit list that is honest about what it cannot do.
$ curl -s "https://emailcheck.ayamk.in/v1/verify?email=tim%40stripe.com" { "validFormat": true, "score": 100, "deliverability": "deliverable", "flags": { "disposable": false, "roleAccount": false, "hasMx": true }, "provider": "Google Workspace / Gmail", "durationMs": 41 }
Everything shipped, nothing retired
Status reflects a real health probe, refreshed every 60 seconds. No vanity numbers.
Narrow scope, verifiable behaviour
One Worker per product
No shared mega-service. A bug in the document parser cannot take down email verification, and each API has its own limits, quotas and release cadence.
Deterministic first
Where a rule beats a model, we use the rule. Keywords, entities, PII, DNS and scoring are code, which means identical input gives identical output and your tests stay meaningful.
Failure is documented
Every endpoint lists its error codes, size caps and timeouts in the OpenAPI document. When something is not supported we say so instead of returning a plausible looking empty response.
Three steps, no sales call
Subscribe on RapidAPI
Pick the free tier, get a key immediately. Billing, invoices and quota enforcement are handled by the marketplace.
Point at the proxy
Send X-RapidAPI-Key and X-RapidAPI-Host to the marketplace URL. Or call the Worker directly with your own key for internal jobs.
Generate a client
Feed /openapi.json into openapi-generator, Postman, Insomnia or an LLM tool call. The spec is the source of truth, not a wiki page.
const res = await fetch(
"https://emailcheck.p.rapidapi.com/v1/verify?email=jane%40acme.com",
{ headers: {
"X-RapidAPI-Key": process.env.RAPIDAPI_KEY,
"X-RapidAPI-Host": "emailcheck.p.rapidapi.com",
}},
);
const data = await res.json();
if (data.deliverability === "undeliverable") reject(data);
What we will not do
No resale of other people's data
Every API here computes something from input you send, or fetches a page you asked about. None of them resell scraped social profiles, private datasets or someone else's licensed feed.
No stored payloads
Documents, emails, page bodies and text are processed in memory inside the request and discarded. What gets logged is the route, status, duration and a request id.
No mystery boxes
If a verdict is a heuristic, the response says so and lists the signals behind it. When we cannot verify something, we return unknown rather than guessing for a nicer looking number.
No dark limits
Size caps, page limits, truncation flags and rate limits are in the spec and in the response. You should never discover a limit in production.
Start on the free tier
Every API has a free plan with a real quota, so you can wire it up and test it before paying anything. Upgrade from the same listing in one click.