Get estimate cost sample
GET
/v1/projects/{id}/cost
Returns the raw page samples the cost estimate is built from — useful if you want to see how the estimate was derived rather than just the final number.
Authorizations
Bearer
JWT bearer token from POST /v1/auth/login.
Type
HTTP (bearer)
or
ApiKey
Scoped API key for non-interactive clients (issued by PBI #1426).
Type
API Key (header: X-Api-Key)
Parameters
Path Parameters
id*
Type
Requiredstring
Format
"uuid"Responses
OK
application/json
Number of RunPod calls the page made (≈ blocks translated + ocr + language).
Format
"int32"Pattern
"^-?(?:0|[1-9]\\d*)$"Total RunPod GPU execution time across the page's OCR/language/translate calls.
Format
"double"Pattern
"^-?(?:0|[1-9]\\d*)(?:\\.\\d+)?(?:[eE][+-]?\\d+)?$"Compute cost at the configured GPU rate.
Format
"double"Pattern
"^-?(?:0|[1-9]\\d*)(?:\\.\\d+)?$"The page this sample is for.
Format
"int32"Pattern
"^-?(?:0|[1-9]\\d*)$"The effective source language this page was translated from (the per-page/project override or the
auto-detected code), lower-cased. Together with TargetLanguageCode it keys the fitted
secondsPerWord per language pair (PBI #1792). Empty when unknown.
Default
""The language this page was translated into (the project's target), lower-cased. Empty when unknown.
Default
""Wall-clock time the page took (includes network + local work).
Format
"double"Pattern
"^-?(?:0|[1-9]\\d*)(?:\\.\\d+)?(?:[eE][+-]?\\d+)?$"How many source words this page carried (its embedded PDF word count, or the OCR'd word count when no
embedded text is available), so the sample is a labeled corpus row for self-calibrating the
word-conditioned rate (PBI #1792). 0 when the word count is unknown — the calibrator ignores it.
Default
0Format
"int32"Pattern
"^-?(?:0|[1-9]\\d*)$"