# Add automation steps

URL: https://docs.glossi.io/repeat-at-scale/automations/steps

Choose what an automation does after its trigger: render with a template, apply AI edits, and export the results.

After the trigger, add steps in the order the work should happen. The editor only offers steps that can follow the one before. Start with **Render** and an export, test that, then add **AI Edit**.

> **Enterprise.** Automations are available on Enterprise.

## Render

Creates a project for each model from a template, then renders it. With a folder trigger, this step also brings each file into Glossi as a model first.

* **Template** (required): a workspace or Glossi template. It sets the scene, lighting, shots and clips.
* **Images**: renders every saved shot. **Resolution** (720p, 1080p, 4K or Custom), **Format** (PNG, JPEG, TIFF) and **Post processing** (on, off, or the workspace default; on uses more usage).
* **Videos**: renders every saved clip. **Resolution** (720p, 1080p, 4K or Custom) and **Quality** (**Fast** or **Best**; Best can take more than twice as long).
* **Variants**: renders each of the model's variants separately.
* **Snap to padding** and **Center framing**: move the camera so the model sits a set number of pixels from the frame edge.

To render the same models in several scenes, add more **Render** steps after the trigger, one per template. Each creates its own project.

## Template Match

Picks the template from the subfolder a file arrived in, instead of one fixed template. Use it after a **Google Drive**, **OneDrive** or **AEM** trigger; subfolders are scanned automatically once this step is added.

1. Tag each template with the name of a subfolder, for example `lifestyle`.
2. Put source files in subfolders with those names.
3. Add **Template Match** and choose under **Tags** which tags it handles, or &#x2A;*All tags (catch-all)**.

Subfolder names match tags without regard to case. Files at the root of the folder, and files in a subfolder no Template Match step claims, are skipped. Render settings are the same as for **Render**.

## AI Edit

Edits every render from the step before. Use brand styles, a prompt, or both:

* **Styles**: pick one or more [AI styles](https://docs.glossi.io/review/ai-styles.md). Each style makes one edited version of every render.
* **AI Prompt**: with styles, the prompt is added to each style. Without styles, it makes one edited version of every render.
* **Group by style**: with styles selected, exports put each style's files in their own folder instead of beside the originals.
* **Lock object position** (on by default): keeps the product's framing and position. Turn it off to let the scene recompose.
* **Check product accuracy**: generates a few options and keeps the one closest to the original product. Uses more usage.
* **Quality**: **Fast**, **Standard**, **High** (2x usage) or **Max** (3x usage).
* **Output**: an image, or a video from each render with a choice of video model and **Duration**. Video is billed per second.

Use **Test this edit** to try it on one of your renders before it runs on everything. For a second prompt, add another **AI Edit** step after this one. A failed edit is retried twice before it is skipped; other edits carry on.

## Export to Google Drive, OneDrive or AEM

Always the last step. It uploads the renders and AI edits to a folder. See [Export results](https://docs.glossi.io/repeat-at-scale/automations/connectors.md#export-results) for accounts, **Export files** and folder layout.

## Material Swap and Create Variants

These steps change each model's materials before it renders.

* **Material Swap**: rules that replace materials on slots whose names contain or match some text. The editor marks this step **Soon**; check with Glossi before relying on it.
* **Create Variants**: a grid of named variants, each assigning a material per slot. Pair it with a **Render** step that has **Variants** on to render every variant.

To apply materials to existing models without an automation, see [Materials across models](https://docs.glossi.io/set-up-the-product/materials-across-models.md).

## What can follow what

| After                           | You can add                            |
| ------------------------------- | -------------------------------------- |
| Manual, New Models              | Material Swap, Create Variants, Render |
| Google Drive, OneDrive, AEM     | Material Swap, Render, Template Match  |
| New Projects                    | Render                                 |
| New Renders                     | AI Edit, Export                        |
| Material Swap                   | Create Variants, Render                |
| Create Variants                 | Render                                 |
| Render, Template Match, AI Edit | AI Edit, Export                        |

## Examples

```text
Google Drive -> Render -> Export to Google Drive
```

```text
OneDrive -> Render -> AI Edit (two styles, Group by style) -> Export to OneDrive
```

```text
New Renders -> AI Edit
```

AI edits every new render in the workspace, except renders from other automations.

```text
Google Drive -> Template Match [lifestyle] -> AI Edit -> Export to Google Drive
             -> Template Match [studio] -> Export to Google Drive
```

Files in `lifestyle/` get AI edits; files in `studio/` are exported as rendered. A catch-all step also picks up files in subfolders that another Template Match step handles, so don't combine it with tagged steps unless you want both.

## Related

* [Choose an automation trigger](https://docs.glossi.io/repeat-at-scale/automations/triggers.md)
* [Connect folders and delivery systems](https://docs.glossi.io/repeat-at-scale/automations/connectors.md)
* [Edit with AI](https://docs.glossi.io/review/edit-with-ai.md)