Create Evaluate

Trigger an evaluation run of a workflow’s committed version through the evaluations pipeline. Evaluate the workflow’s attached dataset, inline data, or a platform dataset id; parameters bind as constant entry inputs on every row. Returns a run id and a results URL to poll or open in the browser.

POST /api/v1/workflows/{id}/evaluate

Python

Python

import requests

url = "https://app.langwatch.ai/api/v1/workflows/{id}/evaluate"

payload = {
    "version_id": "<string>",
    "data": [{}],
    "dataset_id": "<string>",
    "parameters": {},
    "row_indices": [4503599627370495]
}
headers = {
    "X-Auth-Token": "<api-key>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

200

{
  "run_id": "<string>",
  "run_url": "<string>",
  "workflow_version_id": "<string>",
  "version": "<string>"
}

Authorizations

X-Auth-Token · string · header · required

Project API key for sending traces and accessing project-scoped resources. Format: sk-lw-... (no underscore). Obtain one by creating a project via the Admin API or the LangWatch UI.

Path Parameters

id · string · required

Minimum string length: 1

Body

application/json

version_id · string

Committed version to evaluate; defaults to the latest commit

data · object[]

Inline rows to evaluate instead of the workflow's attached dataset

Show child attributes

dataset_id · string

Platform dataset id to evaluate; mutually exclusive with data

parameters · object

Constant entry inputs applied to every row, e.g. a feature flag or PR number

Show child attributes

row_indices · integer[]

Subset of dataset row indices to evaluate

Required range: 0 <= x <= 9007199254740991

Response

200

application/json

Success

run_id · string · required

run_url · string · required

workflow_version_id · string · required

version · string · required