A read API over records you have already validated.
The output schema of a flow becomes a REST endpoint with its own keys, rate limits and pagination contract — so consumers stop calling the source system.
GET/v1/customers
11,204 reqp95 118 mserr 0.02%600/min
GET/v1/customers/:id
4,882 reqp95 96 mserr 0.00%600/min
GET/v1/orders
2,196 reqp95 164 mserr 0.11%300/min
GET/v1/inventory
120 reqp95 212 mserr 0.83%120/min
How it works
Mechanisms, not promises.
01
Generated from the validated output
The schema comes from what survived validation, not from the source system. Consumers never inherit an upstream system’s quirks.
02
Keys are environment-scoped
A key carries its environment in its prefix, is stored hashed, tracks last use, and can be revoked instantly with an optional IP allowlist.
03
Limits that fail politely
Per-key rate limits with X-RateLimit headers on every response and a 429 carrying Retry-After — never a silent drop.
04
A documented pagination contract
Cursor pagination with limit and starting_after, returning data, has_more and next_cursor. The same shape on every collection.
05
Errors you can act on
Every error carries a type, a machine-readable code, the offending parameter and a request id that appears in your logs and in ours.
In the API
The same thing, from a terminal.
curl https://api.zentrixquark.com/v1/customers \ -H "Authorization: Bearer zx_sk_live_…" \ -G -d limit=50 -d starting_after=cus_8f21d0 # → 200 · X-RateLimit-Remaining: 584{ "data": [ { "id": "cus_8f21e2", "email_address": "…" } ], "has_more": true, "next_cursor": "cus_8f21e2"}- Methods
- Read endpoints at launch. Write APIs and GraphQL are on the roadmap, not implemented.
- Explorer
- An in-product explorer that executes against the selected environment using your own key.
- Observability
- Requests, p95 latency and error rate per endpoint, with a request inspector for the last 100 calls.
- Versioning
- A major version in the URL; dated minor changes announced in the changelog.