Suggest example rows to train on
Example input rows to fill the training grid with, so the author is not asked to invent test data for an agent they have not run. Real values are read from one of the org’s own tables when a table’s columns can feed every input; otherwise they are written from the agent’s prompt and the workspace’s AI Context. The set is remembered on the agent, so asking again offers the same rows until regenerate is sent or the agent’s inputs change. Read-only against the workspace: no session, no runs, no credits.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Path Parameters
Body
A request for example rows to fill the training grid with.
How many examples to offer.
1 <= x <= 25Skip the agent's remembered examples and find fresh ones, replacing what was stored. What 'Suggest different examples' asks for; the default replays the set so training the same agent twice offers the same rows.
Response
Successful Response
Example rows to fill the grid with, and where they came from.
How the values were found: read out of one of the org's own tables, or written by the model.
table, generated One set of input values per example, in offer order.
A remembered set was discarded because the agent's inputs no longer matched it, so these are fresh.
These were replayed from the agent rather than found afresh.
The table the values came from; null when they were generated.