Run an agent's evaluations

POST /v1/agents/{agent_id}/evaluation-runs

Rehearse every saved conversation the agent has, against the agent as it is now, and judge each one. Answers at once with the run, queued: read it with GET /v1/evaluation-runs/{id} until its status is finished. Each run is charged as testing. Needs evaluations:run.

Parameters

Name	In	Type	Required	Description
`agent_id`	path	string (uuid)	Yes

Answer

202 with EvaluationRunObject.

Errors

Every error is a problem document with a stable code.

Status	When
`401`	The key is missing, mistyped, unknown, revoked or expired.
`402`	The organization's credit is used up.
`403`	The key lacks a permission, a live key was sent from a browser, or the organization is frozen and the request would change something.
`404`	This key reaches no such agent.
`409`	Saved conversations are already being rehearsed in this organization, or the project is frozen.
`422`	Validation Error
`429`	Too many requests for this key or its organization; see `Retry-After`.
`500`	Something went wrong on our side. Quote the `request_id` to support.

Examples

Run an agent's evaluations

#!/bin/sh
# Run an agent's evaluations. It answers at once. Read the run until its status is finished.
curl -sS --fail-with-body -X POST "https://api.aigently.ai/v1/agents/8a1f3c2e-0b6d-4f5e-9a7c-2d4e6f8a0b1c/evaluation-runs" \
  -H "Authorization: Bearer $AIGENTLY_API_KEY"

Try it

Sends this request to the API from your browser, with a test key. Nothing it does rings a phone.