# Your first generation (/docs/getting-started/first-generation)









Every generation in SimpliGen goes through the same 3-step prompt builder. This walkthrough takes you through all three steps to make your first image.

<Video youtubeId="iI78xBgHFSA" title="Your first generation" />

<Steps>
  <Step>
    ### Choose a media type [#choose-a-media-type]

    When SimpliGen opens, you land on the prompt builder. The first step asks what you want to create. Click **Image** to make a still image, or **Video** to make a clip. For your first generation, choose **Image**.

        <img alt="Step 1 of the prompt builder, choosing Image or Video" src="__img0" />
  </Step>

  <Step>
    ### Pick a style [#pick-a-style]

    Next, choose a **preset**. A preset is a ready-made recipe (a model plus tuned settings) for a particular look or job. Presets are grouped into packs, so open a pack to see the presets inside it, then click the preset you want.

    * Cloud-enabled presets show a **cloud** option and run on our GPUs using credits.
    * Local presets download their models the first time you generate on your own GPU.

    If you are not sure, start with one of the core presets. You can always go back and change it.

        <img alt="Step 2 of the prompt builder, choosing a preset from a pack" src="__img1" />
  </Step>

  <Step>
    ### Describe and generate [#describe-and-generate]

    The final step is where you write your prompt and generate. Type what you want in the prompt box, adjust any settings you like, then click the generate button at the bottom.

    Most settings have a small **i** icon next to them: hover or tap it for a plain-language explanation.

        <img alt="Step 3 of the prompt builder, writing the prompt and generating" src="__img2" />
  </Step>
</Steps>

## Cloud vs local [#cloud-vs-local]

* **Cloud** charges credits per generation and needs no setup. See [Cloud & Credits](/docs/cloud-credits).
* **Local** runs on your machine for free after the one-time model download.

If a preset supports both, click **Generate Cloud** (with a credit cost sub-label) to run on our GPUs, or **Prepare Local** / **Generate Local** to run on your own machine. API-only presets show a single **GENERATE IMAGE** button.

## Where results go [#where-results-go]

When you generate, the job moves to the **Queue drawer**; open it by clicking the **Queue** button on the right rail. The drawer shows active, waiting, and failed jobs. Finished results land in **Gallery** on the left rail. If a generation seems stuck, see [Out of memory / stuck generations](/docs/troubleshooting/out-of-memory).