Start Agent Batch

The agent over every row of a table: one agent run a row, its prompt (and urls) filled from the row's columns as {{column}}, its answer the row's new columns (schema). A few rows run at once (concurrency); each pays for what it uses up to maxCreditsPerRow, and the batch never spends more than rows x that cap, or the balance when that was less -- rows it cannot afford are skipped, free. Rows that cannot run (a bad URL, an empty prompt) are kept as invalid and never charged. 202 with next to poll; the webhook hears agent.batch.finished once.

post/api/v1/agent/batch

Request

  • Base URL: https://api.mesharc.dev (the document declares no server; this is the origin it was published from)
  • URL: https://api.mesharc.dev/api/v1/agent/batch
  • Auth: none declared

Headers

authorizationstring nullable
x-mesharc-tenantstring nullable

Request body

namestring nullable
rowsobject[] nullable
csvstring nullable
promptstring required
urlsstring[] nullable
schemaobject nullable
maxCreditsPerRowinteger nullable
maxStepsinteger nullable
allowedDomainsstring[] nullable
concurrencyinteger nullable
connectionIdstring nullable
loginIdstring nullable
dataConnectionsstring[] nullable

Response

Successful Response

idstring required
kindstring required
namestring required
statusstring required
columnsstring[] required
outputColumnsstring[] required
promptstring required
urlsstring[] required
schemaobject nullable
maxCreditsPerRowinteger required
maxStepsinteger required
domainLimitstring required
allowedDomainsstring[] required
concurrencyinteger required
rowsinteger required
budgetinteger required
creditsUsedinteger required
averagePerRowinteger nullable
modelstring required
nextstring nullable
createdAtstring nullable
finishedAtstring nullable
expiresAtstring nullable
morestring nullable
webhookSecretstring nullable
request_idstring

Changes

No recorded changes to this operation across all 1 revision of this API.