Who may call it, and what it costs
Price. Rows inside the workspace’s own scope: free, up to 1,000,000 a month. Rows outside it: $0.01 a row past the free ones. A public answer older than 90 days costs $0.001 the first time the workspace counts or lists it, and is the workspace’s after, shown in
meta.costCredits. quote: true prices it only and starts nothing. Every reply says what the call cost in meta.costCredits. The prices are on Spending, tracking and stopping.
Arguments
search_answers, search_cited_pages, query_metrics or list_work: the tool whose rows to export
that tool’s arguments, as you would call it (rows: “mentions” or “comparisons”, by: [“question_group”] included); limit is ignored, cursor continues an earlier export
ndjson: one JSON row a line; csv: one column per value, a share as number columns (<name>, <name>_low, <name>_high)
at most this many rows in the file; without it the file holds every row the args return (up to the file’s cap), and a quote counts them min 1, max 20000
GET /v1/tools serves it, every shared object written out; https://agent.parse.gl/openapi.json serves the same schema with the filter and the window named once under components and pointed to by $ref. MCP tools/list leaves out the arguments few calls send, and says the filter and the window once, under search_answers; resolve_ids {"kinds": ["argument"]} lists what it left out, and a call with any of them runs:
Example call
Write a search’s rows to a file. Over REST:- Say what the export would cost, and write nothing:
{"tool":"search_answers","args":{"filter":{"brand":["brd_21e2e46d-5498-46bc-bd54-791c1acd74e7"]}},"format":"ndjson","quote":true}
data, context, next, page and meta; follow next for the calls that come after. A refused call returns { code, message, fix }: see the error table.