Skip to main content
POST
One model call, priced per token, at three grades — with the routing, compression and caching that make the cheap grade cheap.

Authorizations

Authorization
string
header
required

An API key minted in the console, sent as Authorization: Bearer <key>. Keys are project-scoped and carry per-operation scopes. A gck_test_… key runs the full request path — including the real upstream call — and bills nothing.

Every scope is enumerated in x-scopes below and in the RFC 9728 protected-resource metadata at https://cloud.gomry.com/.well-known/oauth-protected-resource, which is also what a 401 points at through its WWW-Authenticate: Bearer resource_metadata="…" header.

Body

application/json
capability
any

Required. cheap | balanced | strong. Chooses the price/quality grade, never a specific model.

system
any

Required. 1–20000 characters of instruction. Treat it as yours: nothing is prepended.

user
any

Required. 1–400000 characters. The content to work on. Long inputs are trimmed to the model's window, head first.

jsonSchema
any

A JSON Schema object. Sent as a constrained-decoding hint where the vendor supports it; always parse defensively, because some ignore it.

maxOutputTokens
any

Caps the answer. 1–16384, and never raised above the capability's own ceiling. Remember that thinking tokens count against it, and check usage.truncated — a call that hits the ceiling still returns 200.

effort
any

minimal | low | default. Lowers the thinking budget on a call that does not need to reason. Ignored by vendors that do not support it.

timeoutMs
any

Milliseconds before the vendor call is abandoned. Can only tighten the capability's own timeout, never extend it.

noCache
any

Skip the result cache for this call. Default false. A cache hit bills the same as a live call — it is our cost that changed, not your result.

Response

Success. usage reports what was metered.

The response is of type object.