The run envelope

The response every synchronous tool returns, around its own result.

The run envelope

All tools that return a run use the same response envelope. The envelope structure and field meanings are consistent across tools; only the result field varies based on the tool. Refer to each tool's page for the structure of its result.

Field	Type	Description
`id`	uuid	This run. Use it to read the run again.
`tool_url_path`	string	The tool that ran, by its slug.
`status`	string	`completed` or `failed`. A synchronous tool has already reached one by the time you get this.
`result`	object \| null	The tool's own answer — the one field whose shape differs per tool. Null on a failed run.
`error`	string \| null	Why the run failed, when it did.
`params`	object	The settings this run used, echoed back — a re-run repeats them.
`credits`	integer	What this run was charged. A failed run is charged nothing.
`latency_ms`	integer \| null	How long the run took.
`input_urls`	object	Download URLs for what you uploaded, keyed by field. Presigned and short-lived — read a run again for fresh ones.
`input_filenames`	object	The filenames you uploaded, under the same keys.
`output_urls`	object	Files the run produced, keyed by name — a masked image, a report. Empty for a tool whose answer is entirely in `result`. Presigned the same way.
`output_filenames`	object	Those files' names, under the same keys.
`created_at`	date-time	When the run started, UTC.