Run one block test, all tests for a single block, or every test in a workflow. Execution is asynchronous: the response returns immediately with aDocumentation Index
Fetch the complete documentation index at: https://docs.retab.com/llms.txt
Use this file to discover all available pages before exploring further.
batch_id + job_id. Poll
jobs.retrieve(job_id) until the job
completes, then read the parsed batch result from job.response.body.
The request body provides EXACTLY ONE of:
test_id — run a single test by id.target — run all tests for a single block ({ type: "block", block_id: ... }).n_consensus is optional and only meaningful for extract / split / classifier
blocks. Allowed values are 3, 5, or 7. Provide it to override the
block’s configured consensus count for this batch only.
jobs.retrieve(job_id) until terminal. When the job completes,
job.response.body is the BlockTestBatchExecutionResult.
failed, cancelled, and expired as terminal failures. If the polling
deadline elapses before a terminal status, raise or return a timeout from your
application code.
response.body containsjobs.retrieve(job_id) returns a
completed job, response.body is the BlockTestBatchExecutionResult:
BlockTestBatchExecutionItem carries the run_record_id you can pass to
Get Block Test Run to fetch the full execution snapshot
(inputs, outputs, assertion_result, verdict_summary, etc.).
The counts object has one bucket per run-record status value.Successful Response
"queued"Public workflow-test target.
The storage layer remains block-scoped today, but the API shape names the tested entity explicitly so workflow-level targets can be added later.