> For the complete documentation index, see [llms.txt](https://docs.umbraco.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.umbraco.com/ai-in-umbraco/17.latest/testing-and-evaluation/api/create.md).

# Create

Creates a new test definition with graders and optional variations.

## Request

```http
POST /umbraco/ai/management/api/v1/tests
```

### Request Body

{% code title="Request" %}

```json
{
    "alias": "test-summarize-quality",
    "name": "Summarization Quality",
    "description": "Validates summarization output quality and format",
    "testFeatureId": "prompt",
    "testTargetId": "d290f1ee-6c54-4b01-90e6-d701748f0851",
    "profileId": "e401f2ff-7d65-5c12-a1f7-e812859a1962",
    "contextIds": [],
    "testFeatureConfig": {
        "variables": {
            "content": "Umbraco is a flexible CMS built on .NET..."
        }
    },
    "graders": [
        {
            "graderTypeId": "contains",
            "name": "Has bullet points",
            "config": { "searchPattern": "- ", "ignoreCase": true },
            "severity": "Error",
            "weight": 1.0
        },
        {
            "graderTypeId": "llm-judge",
            "name": "Quality check",
            "config": {
                "evaluationCriteria": "Is the summary concise and accurate?",
                "passThreshold": 0.7
            },
            "severity": "Error",
            "weight": 1.0
        }
    ],
    "variations": [
        {
            "name": "Claude Sonnet",
            "profileId": "CLAUDE_PROFILE_GUID"
        }
    ],
    "runCount": 3,
    "tags": ["quality", "summarization"]
}
```

{% endcode %}

### Request Properties

| Property            | Type      | Required | Description                               |
| ------------------- | --------- | -------- | ----------------------------------------- |
| `alias`             | string    | Yes      | Unique alias (URL-safe)                   |
| `name`              | string    | Yes      | Display name                              |
| `testFeatureId`     | string    | Yes      | Test feature: `prompt` or `agent`         |
| `testTargetId`      | guid      | Yes      | ID of the prompt or agent to test         |
| `graders`           | array     | Yes      | At least one grader (see below)           |
| `description`       | string    | No       | Optional description                      |
| `profileId`         | guid      | No       | Default AI profile                        |
| `contextIds`        | guid\[]   | No       | Default AI contexts                       |
| `testFeatureConfig` | object    | No       | Feature-specific configuration (JSON)     |
| `variations`        | array     | No       | A/B testing overrides                     |
| `runCount`          | int       | No       | Number of runs per execution (default: 1) |
| `tags`              | string\[] | No       | Organization tags                         |

### Grader Properties

| Property       | Type   | Required | Description                                           |
| -------------- | ------ | -------- | ----------------------------------------------------- |
| `graderTypeId` | string | Yes      | Grader type (for example, `exact-match`, `llm-judge`) |
| `name`         | string | Yes      | Display name for the grader instance                  |
| `description`  | string | No       | What the grader validates                             |
| `config`       | object | No       | Grader-specific configuration (JSON)                  |
| `negate`       | bool   | No       | Invert the result (default: false)                    |
| `severity`     | string | No       | `Info`, `Warning`, or `Error` (default: Error)        |
| `weight`       | double | No       | Weight for scoring, 0 to 1 (default: 1.0)             |

### Variation Properties

| Property            | Type    | Required | Description                           |
| ------------------- | ------- | -------- | ------------------------------------- |
| `name`              | string  | Yes      | Display name                          |
| `description`       | string  | No       | What the variation tests              |
| `profileId`         | guid    | No       | Profile override                      |
| `runCount`          | int     | No       | Run count override                    |
| `contextIds`        | guid\[] | No       | Context IDs override                  |
| `testFeatureConfig` | object  | No       | Feature config override (deep-merged) |

## Response

### Success

Returns `201 Created` with the new test's ID in the response body and a `Location` header pointing to the created test.

{% code title="201 Created" %}

```json
"3fa85f64-5717-4562-b3fc-2c963f66afa6"
```

{% endcode %}

Use [Get Test](/ai-in-umbraco/17.latest/testing-and-evaluation/api/get.md) with the returned ID to retrieve the full test details.

### Validation Error

{% code title="400 Bad Request" %}

```json
{
    "type": "https://tools.ietf.org/html/rfc7231#section-6.5.1",
    "title": "Duplicate alias",
    "status": 400,
    "detail": "A test with this alias already exists."
}
```

{% endcode %}

## Examples

{% code title="cURL" %}

```bash
curl -X POST "https://your-site.com/umbraco/ai/management/api/v1/tests" \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "alias": "test-seo-length",
    "name": "SEO Description Length",
    "testFeatureId": "prompt",
    "testTargetId": "PROMPT_GUID",
    "runCount": 1,
    "graders": [
      {
        "graderTypeId": "regex",
        "name": "Length 50-160 chars",
        "config": { "pattern": "^.{50,160}$" },
        "severity": "Error"
      }
    ],
    "tags": ["seo", "format"]
  }'
```

{% endcode %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.umbraco.com/ai-in-umbraco/17.latest/testing-and-evaluation/api/create.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.