Start an eval run
curl --request POST \
--url https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"dataset_id": "evalDs_2Nk9pQ",
"rubric_name": "correctness",
"rubric_prompt": "Score 0-100 on factual accuracy against the expected answer.",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70
}
'import requests
url = "https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs"
payload = {
"dataset_id": "evalDs_2Nk9pQ",
"rubric_name": "correctness",
"rubric_prompt": "Score 0-100 on factual accuracy against the expected answer.",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({
dataset_id: 'evalDs_2Nk9pQ',
rubric_name: 'correctness',
rubric_prompt: 'Score 0-100 on factual accuracy against the expected answer.',
judge_model: 'claude-haiku-4-5-20251001',
threshold: 70
})
};
fetch('https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs', 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.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs",
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => "",
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => 30,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
CURLOPT_CUSTOMREQUEST => "POST",
CURLOPT_POSTFIELDS => json_encode([
'dataset_id' => 'evalDs_2Nk9pQ',
'rubric_name' => 'correctness',
'rubric_prompt' => 'Score 0-100 on factual accuracy against the expected answer.',
'judge_model' => 'claude-haiku-4-5-20251001',
'threshold' => 70
]),
CURLOPT_HTTPHEADER => [
"Authorization: Bearer <token>",
"Content-Type: application/json"
],
]);
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
echo "cURL Error #:" . $err;
} else {
echo $response;
}package main
import (
"fmt"
"strings"
"net/http"
"io"
)
func main() {
url := "https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs"
payload := strings.NewReader("{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}")
req, _ := http.NewRequest("POST", url, payload)
req.Header.Add("Authorization", "Bearer <token>")
req.Header.Add("Content-Type", "application/json")
res, _ := http.DefaultClient.Do(req)
defer res.Body.Close()
body, _ := io.ReadAll(res.Body)
fmt.Println(string(body))
}HttpResponse<String> response = Unirest.post("https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs")
.header("Authorization", "Bearer <token>")
.header("Content-Type", "application/json")
.body("{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}")
.asString();require 'uri'
require 'net/http'
url = URI("https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs")
http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true
request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}"
response = http.request(request)
puts response.read_body{
"data": {
"id": "<string>",
"agent_id": "<string>",
"dataset_id": "<string>",
"rubric_name": "correctness",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70,
"status": "pending"
},
"meta": {
"request_id": "<string>",
"timestamp": "2023-11-07T05:31:56Z",
"docs_url": "<string>"
}
}Agents
Start an eval run
Launch an LLM-judge eval run that scores the agent against a dataset’s golden rows using the chosen rubric. Pass the dataset_id, a rubric_name (correctness, helpfulness, safety, groundedness, or custom), and optionally a rubric_prompt (required for custom), a Claude judge_model, and a pass threshold (0–100). The run is enqueued and returns immediately with status pending; poll the run by id for scored results.
POST
/
api
/
v1
/
agents
/
{agentId}
/
evals
/
runs
Start an eval run
curl --request POST \
--url https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"dataset_id": "evalDs_2Nk9pQ",
"rubric_name": "correctness",
"rubric_prompt": "Score 0-100 on factual accuracy against the expected answer.",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70
}
'import requests
url = "https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs"
payload = {
"dataset_id": "evalDs_2Nk9pQ",
"rubric_name": "correctness",
"rubric_prompt": "Score 0-100 on factual accuracy against the expected answer.",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({
dataset_id: 'evalDs_2Nk9pQ',
rubric_name: 'correctness',
rubric_prompt: 'Score 0-100 on factual accuracy against the expected answer.',
judge_model: 'claude-haiku-4-5-20251001',
threshold: 70
})
};
fetch('https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs', 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.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs",
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => "",
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => 30,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
CURLOPT_CUSTOMREQUEST => "POST",
CURLOPT_POSTFIELDS => json_encode([
'dataset_id' => 'evalDs_2Nk9pQ',
'rubric_name' => 'correctness',
'rubric_prompt' => 'Score 0-100 on factual accuracy against the expected answer.',
'judge_model' => 'claude-haiku-4-5-20251001',
'threshold' => 70
]),
CURLOPT_HTTPHEADER => [
"Authorization: Bearer <token>",
"Content-Type: application/json"
],
]);
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
echo "cURL Error #:" . $err;
} else {
echo $response;
}package main
import (
"fmt"
"strings"
"net/http"
"io"
)
func main() {
url := "https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs"
payload := strings.NewReader("{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}")
req, _ := http.NewRequest("POST", url, payload)
req.Header.Add("Authorization", "Bearer <token>")
req.Header.Add("Content-Type", "application/json")
res, _ := http.DefaultClient.Do(req)
defer res.Body.Close()
body, _ := io.ReadAll(res.Body)
fmt.Println(string(body))
}HttpResponse<String> response = Unirest.post("https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs")
.header("Authorization", "Bearer <token>")
.header("Content-Type", "application/json")
.body("{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}")
.asString();require 'uri'
require 'net/http'
url = URI("https://api.orbit.devotel.io/api/v1/agents/{agentId}/evals/runs")
http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true
request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n \"dataset_id\": \"evalDs_2Nk9pQ\",\n \"rubric_name\": \"correctness\",\n \"rubric_prompt\": \"Score 0-100 on factual accuracy against the expected answer.\",\n \"judge_model\": \"claude-haiku-4-5-20251001\",\n \"threshold\": 70\n}"
response = http.request(request)
puts response.read_body{
"data": {
"id": "<string>",
"agent_id": "<string>",
"dataset_id": "<string>",
"rubric_name": "correctness",
"judge_model": "claude-haiku-4-5-20251001",
"threshold": 70,
"status": "pending"
},
"meta": {
"request_id": "<string>",
"timestamp": "2023-11-07T05:31:56Z",
"docs_url": "<string>"
}
}Authorizations
BearerApiKey
Dashboard JWT token from Clerk
Path Parameters
Body
application/json
Example:
"evalDs_2Nk9pQ"
Available options:
correctness, helpfulness, safety, groundedness, custom Example:
"correctness"
Required when rubric_name is custom.
Example:
"Score 0-100 on factual accuracy against the expected answer."
Claude model id for the judge. Defaults to the agent's judge model.
Example:
"claude-haiku-4-5-20251001"
Required range:
0 <= x <= 100Example:
70
⌘I