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

# Generate AI-ready data

> Build evaluation datasets from your own source documents.

**AI-Ready Data** builds evaluation datasets from source documents you upload. You describe what the documents contain and what should be generated from them, and SeekrGuard runs the generation as a job you can track from the **AI-Ready Data** page.

<Note>
  AI-Ready Data is in beta. Review a generated dataset before you use it in an examination.
</Note>

## Create a generation job

<Steps>
  <Step title="Create Job">
    On the **AI-Ready Data** page, click **Create Job**.
  </Step>

  <Step title="Define Plan and Upload Data Sources">
    In **Data Generation Plan**, describe the documents you are uploading and what should be generated from them. The plan is required and accepts up to 10,000 characters. Your plan becomes part of the system message on every generated record, so its wording reaches the models you examine later.

    Drag and drop your files onto the upload area, or click **Browse** to choose them from your computer. Upload up to 10 files in PDF, Word, Markdown, or JSON format, each no larger than 100 MB. Your files are listed under **Files Uploaded**, and the **×** on a file removes it.

    Click **Create Job**. The button stays unavailable until you upload at least one data source.
  </Step>
</Steps>

A generated file does not enter your dataset library on its own. To use it in an examination, add it as a dataset. See [Add and manage datasets](/guard/app/datasets).

## Track a job

The **AI-Ready Data Jobs** table lists one row per job, newest first, paginated with **Rows per page**.

| Column              | Contents                                                     |
| ------------------- | ------------------------------------------------------------ |
| **Job ID**          | The job's identifier, prefixed `airdj-`.                     |
| **State**           | Where the job is in its lifecycle.                           |
| **Phase**           | The stage of generation currently running.                   |
| **Created**         | When you submitted the job.                                  |
| **Updated**         | When the job's state last changed.                           |
| **Records**         | The number of records in the generated dataset.              |
| **Eval File ID**    | The generated file's identifier, prefixed `file-`.           |
| **Filename**        | The generated file's name, derived from the source document. |
| **Type**            | The generated file's format.                                 |
| **Stop Reason**     | Why the job stopped short of completing.                     |
| **Failure Summary** | Why the job failed.                                          |