Skip to main content
GET
Get prompt run status
Poll about every 30 seconds until status becomes FINISHED.

FAQ

Why is there no answer body? The answers field in this response only contains the progress of each run. For the body and sources, query List answers with the runID filter. The status is FINISHED but failedCount is non-zero. Even if individual runs fail, the run request ends as FINISHED. Failed runs produce no answer and are not charged.

Authorizations

API-Key
string
header
required

"Enter your organization API key. Example: 'cs_live_xxxx'"

Path Parameters

runID
string
required

ID of the prompt run to retrieve: enter the id from the prompt run request response as is

Response

OK

answers
object[]

Detailed progress status per item (1 item per prompt × model × repeat)

failedCount
integer

Number of failed or skipped items

Example:

12

finishedCount
integer

Number of finished items (success + failed)

Example:

5000

id
string

ID of the retrieved prompt run. Same value as the id in the submit response and the runID in the request path

Example:

"S8cqDSjSot67bCK6YtmvHw"

runningCount
integer

Number of items not yet finished

Example:

0

status
enum<string>

Overall progress status of the prompt run. RUNNING: some items have not finished yet (runningCount > 0), FINISHED: all items have finished (finishedCount == totalCount). It transitions from RUNNING to FINISHED only once and never goes back. Right after submission, while run items are not yet prepared and totalCount is 0, the status is still RUNNING. The status becomes FINISHED once every item has finished even if some failed, so use successCount and failedCount to determine success

Available options:
RUNNING,
FINISHED
Example:

"FINISHED"

successCount
integer

Number of items for which answer classification is also complete

Example:

4988

totalCount
integer

Total number of run items (number of prompts × number of models × repeat count)

Example:

5000