API reference

Endpoints

10 endpoints, all under https://api.cresva.ai/v1. Read only: nothing in this API changes a campaign, a budget or a creative on any platform.

GET/v1/brandsList the brands this key can reach GET/v1/metricsAd performance, at any level POST/v1/queryAsk in plain language GET/v1/forecastsStored forecasts with confidence intervals GET/v1/anomaliesDetected outliers in performance GET/v1/attributionAttributed performance under your own model GET/v1/creativesPer-creative performance POST/v1/recommendationsBudget reallocation under constraints GET/v1/reportsReports this brand has generated GET/v1/syncHow fresh your data is

Conventions

Two of these are POST and neither of them writes anything. /v1/query and /v1/recommendations take a body because their inputs do not fit comfortably in a query string, not because they mutate state.

Every response carries a meta object with a request_id and a simulated flag. Rate-limit headers are on every response, including errors.

Generate a client instead

The specification these pages are built from is served at /v1/openapi.json. Point a generator at it and you get a typed client in your own language rather than hand-written request code.

Spec

curl https://api.cresva.ai/v1/openapi.json -o cresva-openapi.json

Conventions#

Conventions