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

# Manage datasets

The `eval dataset` command group manages evaluation sets: list, show, create, and delete datasets, and add or remove cases within a dataset. A dataset is the data source for an experiment; its fields (such as `input` and `reference_output`) are mapped to the target and the evaluator during [`eval run`](/productions/agentkit-cli/archives/0.50.9/en/commands/eval/run).

<Note>
  These subcommands are also available as top-level aliases, so the `eval` prefix may be omitted: `agentkit eval dataset list` is equivalent to `agentkit dataset list`. The examples below use the shorter `agentkit dataset ...` form. All subcommands share `-p, --project <name>` (project name, default `default`) and `-r, --region <region>` (Volcengine region; default auto-detect, or from env for `create`). Most also accept `--json` (output raw JSON); `dataset delete` does not.
</Note>

## dataset list

List evaluation datasets.

```bash lines theme={null}
agentkit dataset list
```

## dataset show

Show details of an evaluation dataset, including its case items.

| Flag / Argument | Description | Default |
| - | - | - |
| `<id>` | Evaluation set id or name (required) | None |
| `--items <n>` | List up to N case items (0 to skip) | `20` |

```bash lines theme={null}
agentkit dataset show qa-set --items 50
```

## dataset create

Create an evaluation dataset. The schema is fixed once created, and a case's keys must match it.

| Flag / Argument | Description | Default |
| - | - | - |
| `--name <name>` | Dataset name (required) | None |
| `--schema <fields>` | Comma-separated field names | `input,reference_output,output` |
| `--description <text>` | Dataset description | None |

```bash lines theme={null}
agentkit dataset create --name qa-set --schema "input,reference_output"
```

<Tip>
  `output` (the model's actual output) is usually produced by the evaluated target runtime during the experiment, so it does **not** need to be part of the dataset schema. A dataset typically only needs `input` and `reference_output`.
</Tip>

## dataset add

Add one or more cases to an evaluation dataset.

| Flag / Argument | Description | Default |
| - | - | - |
| `<id>` | Evaluation set id or name (required) | None |
| `--field <key=value>` | Case field (repeatable) | None |
| `--file <path>` | JSON file: an object, or an array of `{field: value}` objects | None |

```bash lines theme={null}
agentkit dataset add qa-set --field "input=What is the capital of France?" --field "reference_output=Paris"
```

## dataset remove

Remove one or more cases from an evaluation dataset.

| Flag / Argument | Description | Default |
| - | - | - |
| `<id>` | Evaluation set id or name (required) | None |
| `<itemIds...>` | Case item id(s) to remove (one or more, required) | None |
| `-y, --yes` | Skip the confirmation prompt | `false` |

```bash lines theme={null}
agentkit dataset remove qa-set item-1 item-2 -y
```

## dataset delete

Delete an evaluation dataset.

| Flag / Argument | Description | Default |
| - | - | - |
| `<id>` | Evaluation set id or name (required) | None |
| `-y, --yes` | Skip the confirmation prompt | `false` |

```bash lines theme={null}
agentkit dataset delete qa-set -y
```