Expand dataset with synthetic samples

POST/v1/private/datasets/{id}/expansionsExpand dataset with synthetic samples

Generate synthetic dataset samples using LLM based on existing data patterns

Parameters

idstring · uuidpathrequired

Request body

application/json

objectDatasetExpansion_Write

custom_promptstring

Custom prompt to use for generation instead of auto-generated one

max_completion_tokensinteger · int32

Maximum number of tokens for the LLM response. Required by Anthropic, used as maxOutputTokens for Gemini. If not provided, defaults to 4000 for Anthropic models only.

minimum 100

modelstringrequired

The model to use for synthetic data generation

minLength 1

preserve_fieldsarray of string

Fields to preserve patterns from original data

Show child attributes

sample_countinteger · int32

Number of synthetic samples to generate

maximum 200 · minimum 1

variation_instructionsstring

Additional instructions for data variation

Example request

{
  "custom_prompt": "string",
  "max_completion_tokens": 0,
  "model": "gpt-4",
  "preserve_fields": [
    "input",
    "expected_output"
  ],
  "sample_count": 10,
  "variation_instructions": "Create variations that test edge cases"
}

Responses

objectDatasetExpansionResponse

generated_samplesarray of object

List of generated synthetic dataset items

Show child attributes

Show array items

created_atstring · date-time

created_bystring

dataobjectrequired

dataset_idstring · uuid

dataset_item_idstring · uuid

Deprecated. Always equals 'id'. Retained for backward compatibility and will be removed in a future version

descriptionstring

pattern (?s)^\s*(\S.*\S|\S)\s*$

evaluatorsarray of object

Show child attributes

Show array items

configobjectrequired

namestringrequired

minLength 1

typestringrequired

one of "llm_judge", "code_metric"

execution_policyobject

Show child attributes

pass_thresholdinteger · int32

maximum 100 · minimum 1

runs_per_iteminteger · int32

maximum 100 · minimum 1

experiment_itemsarray of object

Show child attributes

Show array items

assertion_resultsarray of object

Show child attributes

Show array items

passedboolean

reasonstring

valuestring

commentsarray of object

Show child attributes

Show array items

created_atstring · date-time

created_bystring

idstring · uuid

last_updated_atstring · date-time

last_updated_bystring

source_queue_idstring · uuid

textstringrequired

minLength 1

created_atstring · date-time

created_bystring

dataset_item_idstring · uuidrequired

descriptionstring

durationnumber · double

execution_policyobjectExecutionPolicy ↑

experiment_idstring · uuidrequired

feedback_scoresarray of object

Show child attributes

Show array items

category_namestring

created_atstring · date-time

created_bystring

last_updated_atstring · date-time

last_updated_bystring

namestringrequired

minLength 1

reasonstring

sourcestringrequired

one of "ui", "sdk", "online_scoring"

source_queue_idstring · uuid

valuenumberrequired

maximum 1000000000 · minimum -1000000000

value_by_authorobject

idstring · uuid

inputobject

Show child attributes

anyOf · 3 options

Option 1object

Option 2array of object

Option 3string

last_updated_atstring · date-time

last_updated_bystring

outputobjectJsonListString ↑

project_idstring · uuid

project_namestring

pattern (?s)^\s*(\S.*\S|\S)\s*$

statusstring

one of "passed", "failed"

total_estimated_costnumber

trace_idstring · uuidrequired

trace_metadataobjectJsonListString ↑

trace_visibility_modestring

one of "default", "hidden"

usageobject

idstring · uuid

Stable item identifier. On write, used as the upsert key. If omitted, a new ID is generated. Remains the same across dataset versions

last_updated_atstring · date-time

last_updated_bystring

run_summaries_by_experimentobject

sourcestringrequired

one of "manual", "trace", "span", "sdk"

span_idstring · uuid

tagsarray of string

uniqueItems true

Show child attributes

uniqueItems true

trace_idstring · uuid

generation_timestring · date-time

Generation timestamp

modelstring

Model used for generation

total_generatedinteger · int32

Total number of samples generated

Example response

{
  "generated_samples": [
    {
      "created_at": "2026-06-09T00:00:00Z",
      "created_by": "string",
      "data": {},
      "dataset_id": "00000000-0000-0000-0000-000000000000",
      "dataset_item_id": "00000000-0000-0000-0000-000000000000",
      "description": "string",
      "evaluators": [
        {
          "config": {},
          "name": "string",
          "type": "code_metric"
        }
      ],
      "execution_policy": {
        "pass_threshold": 0,
        "runs_per_item": 0
      },
      "experiment_items": [
        {
          "assertion_results": [
            {
              "passed": true,
              "reason": "string",
              "value": "string"
            }
          ],
          "comments": [
            {
              "created_at": "2026-06-09T00:00:00Z",
              "created_by": "string",
              "id": "00000000-0000-0000-0000-000000000000",
              "last_updated_at": "2026-06-09T00:00:00Z",
              "last_updated_by": "string",
              "source_queue_id": "00000000-0000-0000-0000-000000000000",
              "text": "string"
            }
          ],
          "created_at": "2026-06-09T00:00:00Z",
          "created_by": "string",
          "dataset_item_id": "00000000-0000-0000-0000-000000000000",
          "description": "string",
          "duration": 0,
          "execution_policy": {
            "pass_threshold": 0,
            "runs_per_item": 0
          },
          "experiment_id": "00000000-0000-0000-0000-000000000000",
          "feedback_scores": [
            {
              "category_name": "string",
              "created_at": "2026-06-09T00:00:00Z",
              "created_by": "string",
              "last_updated_at": "2026-06-09T00:00:00Z",
              "last_updated_by": "string",
              "name": "string",
              "reason": "string",
              "source": "online_scoring",
              "source_queue_id": "00000000-0000-0000-0000-000000000000",
              "value": 0,
              "value_by_author": {
                "additionalProp1": {
                  "author": "string",
                  "category_name": "string",
                  "last_updated_at": "2026-06-09T00:00:00Z",
                  "reason": "string",
                  "source": "online_scoring",
                  "source_queue_id": "string",
                  "span_id": "string",
                  "span_type": "string",
                  "value": 0
                }
              }
            }
          ],
          "id": "00000000-0000-0000-0000-000000000000",
          "input": [
            {}
          ],
          "last_updated_at": "2026-06-09T00:00:00Z",
          "last_updated_by": "string",
          "output": [
            {}
          ],
          "project_id": "00000000-0000-0000-0000-000000000000",
          "project_name": "string",
          "status": "failed",
          "total_estimated_cost": 0,
          "trace_id": "00000000-0000-0000-0000-000000000000",
          "trace_metadata": [
            {}
          ],
          "trace_visibility_mode": "default",
          "usage": {
            "additionalProp1": 0
          }
        }
      ],
      "id": "00000000-0000-0000-0000-000000000000",
      "last_updated_at": "2026-06-09T00:00:00Z",
      "last_updated_by": "string",
      "run_summaries_by_experiment": {
        "additionalProp1": {
          "passed_runs": 0
        }
      },
      "source": "manual"
    }
  ]
}

Examples

idrequiredpath parameter

curl -X POST 'http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions' \
  -H 'Content-Type: application/json' \
  -d '{"custom_prompt":"string","max_completion_tokens":0,"model":"gpt-4","preserve_fields":["input","expected_output"],"sample_count":10,"variation_instructions":"Create variations that test edge cases"}'

const response = await fetch('http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({"custom_prompt":"string","max_completion_tokens":0,"model":"gpt-4","preserve_fields":["input","expected_output"],"sample_count":10,"variation_instructions":"Create variations that test edge cases"}),
});
const data = await response.json();

const response = await fetch('http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions', { method: 'POST', headers: { 'Content-Type': 'application/json' }, body: "{\"custom_prompt\":\"string\",\"max_completion_tokens\":0,\"model\":\"gpt-4\",\"preserve_fields\":[\"input\",\"expected_output\"],\"sample_count\":10,\"variation_instructions\":\"Create variations that test edge cases\"}" });
console.log(await response.json());

import requests

response = requests.request(
    "POST",
    "http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions",
    headers={"Content-Type": "application/json"},
    data='{"custom_prompt":"string","max_completion_tokens":0,"model":"gpt-4","preserve_fields":["input","expected_output"],"sample_count":10,"variation_instructions":"Create variations that test edge cases"}',
)
print(response.json())

package main

import (
	"context"
	"net/http"
	"strings"
)

func main() {
	ctx := context.Background()
	body := strings.NewReader(`{"custom_prompt":"string","max_completion_tokens":0,"model":"gpt-4","preserve_fields":["input","expected_output"],"sample_count":10,"variation_instructions":"Create variations that test edge cases"}`)
	req, err := http.NewRequestWithContext(ctx, "POST", "http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions", body)
	if err != nil {
		panic(err)
	}
	req.Header.Set("Content-Type", "application/json")
	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		panic(err)
	}
	defer resp.Body.Close()
}

<?php
$request = curl_init("http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions");
curl_setopt($request, CURLOPT_CUSTOMREQUEST, "POST");
curl_setopt($request, CURLOPT_POSTFIELDS, "{\"custom_prompt\":\"string\",\"max_completion_tokens\":0,\"model\":\"gpt-4\",\"preserve_fields\":[\"input\",\"expected_output\"],\"sample_count\":10,\"variation_instructions\":\"Create variations that test edge cases\"}");
curl_setopt($request, CURLOPT_HTTPHEADER, ['Content-Type: application/json']);
$response = curl_exec($request);
curl_close($request);

var request = HttpRequest.newBuilder(URI.create("http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions"))
    .method("POST", HttpRequest.BodyPublishers.noBody())
    .header("Content-Type", "application/json")
    .method("POST", HttpRequest.BodyPublishers.ofString("{\"custom_prompt\":\"string\",\"max_completion_tokens\":0,\"model\":\"gpt-4\",\"preserve_fields\":[\"input\",\"expected_output\"],\"sample_count\":10,\"variation_instructions\":\"Create variations that test edge cases\"}"))
    .build();
var response = HttpClient.newHttpClient().send(request, HttpResponse.BodyHandlers.ofString());

require 'net/http'

uri = URI("http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions")
request = Net::HTTPGenericRequest.new("POST", true, true, uri)
request['Content-Type'] = 'application/json'
request.body = "{\"custom_prompt\":\"string\",\"max_completion_tokens\":0,\"model\":\"gpt-4\",\"preserve_fields\":[\"input\",\"expected_output\"],\"sample_count\":10,\"variation_instructions\":\"Create variations that test edge cases\"}"
response = Net::HTTP.start(uri.hostname, uri.port, use_ssl: true) { |http| http.request(request) }

using var client = new HttpClient();
using var request = new HttpRequestMessage(HttpMethod.Post, "http://localhost:5173/api/v1/private/datasets/00000000-0000-0000-0000-000000000000/expansions");
request.Content = new StringContent("{\"custom_prompt\":\"string\",\"max_completion_tokens\":0,\"model\":\"gpt-4\",\"preserve_fields\":[\"input\",\"expected_output\"],\"sample_count\":10,\"variation_instructions\":\"Create variations that test edge cases\"}", Encoding.UTF8, "application/json");
using var response = await client.SendAsync(request);

{
  "generated_samples": [
    {
      "created_at": "2026-06-09T00:00:00Z",
      "created_by": "string",
      "data": {},
      "dataset_id": "00000000-0000-0000-0000-000000000000",
      "dataset_item_id": "00000000-0000-0000-0000-000000000000",
      "description": "string",
      "evaluators": [
        {
          "config": {},
          "name": "string",
          "type": "code_metric"
        }
      ],
      "execution_policy": {
        "pass_threshold": 0,
        "runs_per_item": 0
      },
      "experiment_items": [
        {
          "assertion_results": [
            {
              "passed": true,
              "reason": "string",
              "value": "string"
            }
          ],
          "comments": [
            {
              "created_at": "2026-06-09T00:00:00Z",
              "created_by": "string",
              "id": "00000000-0000-0000-0000-000000000000",
              "last_updated_at": "2026-06-09T00:00:00Z",
              "last_updated_by": "string",
              "source_queue_id": "00000000-0000-0000-0000-000000000000",
              "text": "string"
            }
          ],
          "created_at": "2026-06-09T00:00:00Z",
          "created_by": "string",
          "dataset_item_id": "00000000-0000-0000-0000-000000000000",
          "description": "string",
          "duration": 0,
          "execution_policy": {
            "pass_threshold": 0,
            "runs_per_item": 0
          },
          "experiment_id": "00000000-0000-0000-0000-000000000000",
          "feedback_scores": [
            {
              "category_name": "string",
              "created_at": "2026-06-09T00:00:00Z",
              "created_by": "string",
              "last_updated_at": "2026-06-09T00:00:00Z",
              "last_updated_by": "string",
              "name": "string",
              "reason": "string",
              "source": "online_scoring",
              "source_queue_id": "00000000-0000-0000-0000-000000000000",
              "value": 0,
              "value_by_author": {
                "additionalProp1": {
                  "author": "string",
                  "category_name": "string",
                  "last_updated_at": "2026-06-09T00:00:00Z",
                  "reason": "string",
                  "source": "online_scoring",
                  "source_queue_id": "string",
                  "span_id": "string",
                  "span_type": "string",
                  "value": 0
                }
              }
            }
          ],
          "id": "00000000-0000-0000-0000-000000000000",
          "input": [
            {}
          ],
          "last_updated_at": "2026-06-09T00:00:00Z",
          "last_updated_by": "string",
          "output": [
            {}
          ],
          "project_id": "00000000-0000-0000-0000-000000000000",
          "project_name": "string",
          "status": "failed",
          "total_estimated_cost": 0,
          "trace_id": "00000000-0000-0000-0000-000000000000",
          "trace_metadata": [
            {}
          ],
          "trace_visibility_mode": "default",
          "usage": {
            "additionalProp1": 0
          }
        }
      ],
      "id": "00000000-0000-0000-0000-000000000000",
      "last_updated_at": "2026-06-09T00:00:00Z",
      "last_updated_by": "string",
      "run_summaries_by_experiment": {
        "additionalProp1": {
          "passed_runs": 0
        }
      },
      "source": "manual"
    }
  ]
}