# Generate Image

Create an image from a written description.

## What you give it

### `prompt`

**Required**

Describe the image you want, in as much detail as you can.

For example `A serene mountain lake at sunset`

### `model`

**Required**

One of: `GPT Image 1`, `DALL-E 3`, `Playground v3`, `flux-kontext-pro`, `flux-kontext-max`, `flux-1.1-pro`, `Ideogram`, `Recraft`

### `model_style`

**Required**

One of: `default`, `photo`, `digital art`, `illustration`, `drawing`

### `model_aspect_ratio`

**Required**

One of: `9:16`, `1:1`, `4:3`, `16:9`

Any of these can be a value you type, a CRM property token, or the output of an earlier step.

## What you get back

An image hosted for you, returned inside an img tag rather than as a bare URL.

A real response, shortened:

```
<img src="https://…/a-serene-mountain-lake-at-sunset.jpg" alt="A serene mountain lake at sunset" />
```

## Worth knowing

> Because the response is an img tag, extract the URL from its src attribute before passing it to a step that expects a plain link.

[Install in HubSpot](https://builderpack.agent.ai/install)

---

Technical details — action name `generate_image`, type Generate, workflow action ID 1-268463153.