List immutable evaluation verdicts

curl --request GET \
  --url https://api.recursion.labelbox.com/managed-agents/v1/evaluations \
  --header 'Authorization: Bearer <token>'

import requests

url = "https://api.recursion.labelbox.com/managed-agents/v1/evaluations"

headers = {"Authorization": "Bearer <token>"}

response = requests.get(url, headers=headers)

print(response.text)

const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.recursion.labelbox.com/managed-agents/v1/evaluations', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.recursion.labelbox.com/managed-agents/v1/evaluations",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://api.recursion.labelbox.com/managed-agents/v1/evaluations"

	req, _ := http.NewRequest("GET", url, nil)

	req.Header.Add("Authorization", "Bearer <token>")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

require 'uri'
require 'net/http'

url = URI("https://api.recursion.labelbox.com/managed-agents/v1/evaluations")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Get.new(url)
request["Authorization"] = 'Bearer <token>'

response = http.request(request)
puts response.read_body

{
  "evaluations": [
    {
      "archetype": "artifact",
      "child_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "created_at": "2023-11-07T05:31:56Z",
      "criteria": [
        {
          "criterion_key": "<string>",
          "verdict": "pass",
          "criterionText": "<string>",
          "evidence_event_ids": [
            "3c90c3cc-0d44-4b50-8888-8dd25736052a"
          ],
          "label": "<string>",
          "rationale": "<string>"
        }
      ],
      "evaluation_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluationBatchId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluationSessionId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluator_agent_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluator_agent_version_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "failure_class": "none",
      "result": "pass",
      "run_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "size_bucket": "unknown",
      "snapshot_event_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_agent_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_agent_version_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "foldedAt": "2023-11-07T05:31:56Z",
      "outcomeId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "outcomeIteration": 4503599627370495,
      "outcomeResult": "satisfied",
      "summary": "<string>",
      "trigger": "outcome",
      "writerAction": "revise"
    }
  ],
  "next_page_token": "<string>"
}

{
  "code": "invalid_request",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "unauthorized",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "forbidden",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "not_found",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "payload_too_large",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "rate_limit_exceeded",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "internal_error",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "bad_gateway",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "managed_agents_unavailable",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "gateway_timeout",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

Evaluations

List immutable evaluation verdicts

Returns an organization-scoped page of immutable evaluator verdicts ordered by created_at descending, then evaluation_id ascending. Optional filters are combined with AND; createdAtGte and createdAtLt select the half-open interval [createdAtGte, createdAtLt). Retained same-organization evaluation history remains readable after referenced sessions or agents are deleted; unknown and cross-organization session, agent, version, and run references are indistinguishable 404 responses.

GET

managed-agents

evaluations

curl --request GET \
  --url https://api.recursion.labelbox.com/managed-agents/v1/evaluations \
  --header 'Authorization: Bearer <token>'

import requests

url = "https://api.recursion.labelbox.com/managed-agents/v1/evaluations"

headers = {"Authorization": "Bearer <token>"}

response = requests.get(url, headers=headers)

print(response.text)

const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.recursion.labelbox.com/managed-agents/v1/evaluations', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.recursion.labelbox.com/managed-agents/v1/evaluations",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://api.recursion.labelbox.com/managed-agents/v1/evaluations"

	req, _ := http.NewRequest("GET", url, nil)

	req.Header.Add("Authorization", "Bearer <token>")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

require 'uri'
require 'net/http'

url = URI("https://api.recursion.labelbox.com/managed-agents/v1/evaluations")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Get.new(url)
request["Authorization"] = 'Bearer <token>'

response = http.request(request)
puts response.read_body

{
  "evaluations": [
    {
      "archetype": "artifact",
      "child_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "created_at": "2023-11-07T05:31:56Z",
      "criteria": [
        {
          "criterion_key": "<string>",
          "verdict": "pass",
          "criterionText": "<string>",
          "evidence_event_ids": [
            "3c90c3cc-0d44-4b50-8888-8dd25736052a"
          ],
          "label": "<string>",
          "rationale": "<string>"
        }
      ],
      "evaluation_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluationBatchId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluationSessionId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluator_agent_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "evaluator_agent_version_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "failure_class": "none",
      "result": "pass",
      "run_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "size_bucket": "unknown",
      "snapshot_event_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_agent_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_agent_version_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "target_session_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "foldedAt": "2023-11-07T05:31:56Z",
      "outcomeId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "outcomeIteration": 4503599627370495,
      "outcomeResult": "satisfied",
      "summary": "<string>",
      "trigger": "outcome",
      "writerAction": "revise"
    }
  ],
  "next_page_token": "<string>"
}

{
  "code": "invalid_request",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "unauthorized",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "forbidden",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "not_found",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "payload_too_large",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "rate_limit_exceeded",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "internal_error",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "bad_gateway",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "managed_agents_unavailable",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

{
  "code": "gateway_timeout",
  "message": "<string>",
  "details": {
    "field": "<string>",
    "issues": [
      "<string>"
    ],
    "requestId": "<string>",
    "retryable": true
  }
}

Authorizations

Authorization

string

header

required

A Recursion API key, created in the console under API keys.

Headers

x-tenant-id

string

Optional tenant consistency check. When present, it must name the tenant bound to the API key, including when x-organization-id is default.

Minimum string length: 1

x-organization-id

string

Organization in which to act. Required for a tenant-scoped API key. Optional for an organization-scoped key, where it must name that same organization. The value may be an organization id or default.

Minimum string length: 1

Query Parameters

target_session_id

string<uuid>

Narrow to one target session. Unknown, deleted-without-history, and cross-organization ids return 404.

target_agent_id

string<uuid>

Narrow to verdicts for one target agent. Retained same-organization history remains readable after deletion.

evaluator_agent_id

string<uuid>

Narrow to verdicts produced by one evaluator agent. Retained same-organization history remains readable after deletion.

run_session_id

string<uuid>

Narrow to one evaluation run, by the run_session_id managedAgentsListEvaluationRuns returns.

limit

integer

default:50

Maximum verdicts to return. Defaults to 50. The schema maximum is the summary view's: view=summary accepts up to 1000, while view=full (the default) accepts up to 100 and answers a larger limit with a 400.

Required range: 1 <= x <= 1000

page_token

string

Opaque signed continuation from next_page_token. It is bound to this organization, filters, and effective limit and expires after one hour.

targetAgentVersionId

string<uuid>

Narrow to verdicts for one immutable target-agent version. Requires evaluatorAgentVersionId. A version neither in the organization's catalog nor judged by that evaluator version returns 404.

evaluatorAgentVersionId

string<uuid>

Narrow to verdicts produced by one immutable evaluator version. A version neither in the organization's catalog nor in its retained verdicts returns 404.

createdAtGte

string<date-time>

Only verdicts recorded at or after this RFC 3339 instant.

createdAtLt

string<date-time>

Only verdicts recorded before this RFC 3339 instant. Must be after createdAtGte when both are supplied.

failedCriterionKey

string

Only verdicts whose rubric criterion with this key failed. Requires evaluatorAgentVersionId.

Maximum string length: 64

Pattern: ^[a-z][a-z0-9]*(?:[-_][a-z0-9]+)*$

view

enum<string>

default:full

full (default) returns every verdict field. summary omits summary and each criterion's rationale and evidence_event_ids — the evaluator's free text, most of a row's bytes — for callers that only aggregate verdicts. It shapes the representation only, so a page_token continues under either view when its limit is one both views accept (100 or less).

Available options:

full,

summary

trigger

enum<string>

Only verdicts of one trigger: outcome, a grading pass of a writer's outcome, or manual, an evaluation a caller started. Omit for manual evaluations and, once outcome history is unified, outcome rounds too; before then, a list with no target_session_id or run_session_id holds manual evaluations only. A historical outcome round whose grader is no longer recorded is listed only in its outcome's evaluations.

Available options:

outcome,

manual

outcomeId

string<uuid>

Only the verdicts of one outcome, newest round first. Requires target_session_id, the outcome's writer session. A historical round whose grader is no longer recorded is listed only in the outcome's evaluations.

Response

One organization-scoped page of immutable evaluation verdicts and an optional continuation.

evaluations

object[]

required

Newest-first immutable evaluation verdicts for this page.

Show child attributes

next_page_token

string

Signed one-hour continuation bound to the organization, filters, and effective limit.

API reference

Endpoints

List immutable evaluation verdicts

Authorizations

Headers

Query Parameters

Response