> ## Documentation Index
> Fetch the complete documentation index at: https://www.adaline.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompt evaluators

# PromptEvaluatorsClient

`adaline.prompts.evaluators` manages the evaluators attached to a prompt — LLM-as-a-judge graders, JavaScript checks, text matchers, cost, latency, and response-length guards. Evaluators are always scoped to a prompt; there is no workspace-level evaluators collection. Every method is async.

## Access

```python theme={null}
from adaline.main import Adaline

adaline = Adaline()
evaluators = adaline.prompts.evaluators  # PromptEvaluatorsClient
```

The class is also exported directly:

```python theme={null}
from adaline.clients import PromptEvaluatorsClient
```

Types from `adaline_api`:

```python theme={null}
from adaline_api.models.evaluator import Evaluator
from adaline_api.models.create_evaluator_request import CreateEvaluatorRequest
from adaline_api.models.update_evaluator_request import UpdateEvaluatorRequest
from adaline_api.models.list_evaluators_response import ListEvaluatorsResponse
```

Evaluator types at a glance:

| `type`            | What it measures                      |
| ----------------- | ------------------------------------- |
| `llm-as-a-judge`  | Qualitative grading via an LLM rubric |
| `javascript`      | Arbitrary JS/TS sandboxed check       |
| `text-matcher`    | String contains / regex / equality    |
| `cost`            | Cost threshold per row                |
| `latency`         | Response time threshold               |
| `response-length` | Token / character bounds              |

***

## list()

List evaluators attached to a prompt (paginated).

```python theme={null}
async def list(
    *,
    prompt_id: str,
    limit: Optional[int] = None,
    cursor: Optional[str] = None,
    sort: Optional[SortOrderInput] = None,
    created_after: Optional[int] = None,
    created_before: Optional[int] = None,
) -> ListEvaluatorsResponse
```

### Example

```python theme={null}
response = await adaline.prompts.evaluators.list(
    prompt_id="prompt_abc123",
    limit=50,
)
```

***

## create()

Attach a new evaluator to a prompt.

```python theme={null}
async def create(
    *,
    prompt_id: str,
    evaluator: CreateEvaluatorRequest,
) -> Evaluator
```

### Example — LLM-as-a-judge

```python theme={null}
from adaline_api.models.create_evaluator_request import CreateEvaluatorRequest

evaluator = await adaline.prompts.evaluators.create(
    prompt_id="prompt_abc123",
    evaluator=CreateEvaluatorRequest(
        type="llm-as-a-judge",
        title="Factuality",
        settings={
            "model": "gpt-4o",
            "rubric": "Rate 1-5 for factual accuracy against the reference answer.",
            "threshold": 4,
        },
    )
)
```

***

## get()

Fetch a single evaluator by ID.

```python theme={null}
async def get(*, prompt_id: str, evaluator_id: str) -> Evaluator
```

***

## update()

Update an evaluator's title, settings, or threshold.

```python theme={null}
async def update(
    *,
    prompt_id: str,
    evaluator_id: str,
    evaluator: UpdateEvaluatorRequest,
) -> Evaluator
```

### Example

```python theme={null}
from adaline_api.models.update_evaluator_request import UpdateEvaluatorRequest

await adaline.prompts.evaluators.update(
    prompt_id="prompt_abc123",
    evaluator_id="evaluator_abc123",
    evaluator=UpdateEvaluatorRequest(settings={"threshold": 3}),
)
```

***

## delete()

Permanently delete an evaluator.

```python theme={null}
async def delete(*, prompt_id: str, evaluator_id: str) -> None
```

***

## See Also

* [PromptsClient](/docs/reference/sdk/v2/python/classes/prompts) — parent client
* [PromptEvaluationsClient](/docs/reference/sdk/v2/python/classes/prompt-evaluations)
* API reference: [List evaluators](/docs/reference/api/v2/openapi/list-evaluators) · [Create](/docs/reference/api/v2/openapi/create-evaluator) · [Update](/docs/reference/api/v2/openapi/update-evaluator) · [Delete](/docs/reference/api/v2/openapi/delete-evaluator)