Retrieve a run
Reads one run from stored state. Requires runs_read. Never calls the model, so reading a run is free.
A completed run carries its verdict or decision, plus usage and cost_micros. Stored reads omit cost_basis and schema_enforcement, and a stored decision omits model, stages and the thought fields. A run that is not completed returns the run without an answer.
Authorizations
An API key (otsk_ + 64 hex) or a console session token.
Path Parameters
run_ followed by 32 hex characters, as the API returns it. The hyphenated form the usage routes return is also accepted.
Response
The run as stored.
A run. A completed run carries verdict or decision, never both. Absent optional fields are omitted, not null. Stored reads (replays, retrieve, stream, list rows) omit cost_basis and schema_enforcement; list rows also omit the answer, usage and cost.
run_ followed by 32 hex characters.
Which kind of answer this run produces: look at verdict or at decision.
verdict, decision pending, running, completed, failed 64-character hex SHA-256 of the normalized input, kind and contract.
true whenever the body came from storage: every replay, retrieve, stream and list row. A replay is never charged again.
Completed verdict runs: your JSON document, validated against your schema.
The decision's answer set. Present on a completed decision run.
Tokens the run used.
x >= 0Whether cost_micros was reported by the model service or estimated from the price list. Present exactly when cost_micros is.
provider_reported, estimated What constrained the verdict. Present on a fresh verdict run; absent on a replay and on a list row.
unconstrained, requested, forced, enforced