> For the complete documentation index, see [llms.txt](https://docs.ideogram.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ideogram.ai/create/create.md).

# Create Images

Generate AI images with Ideogram from text or reference photos. Choose a model, check credit costs, find results and troubleshoot a generation that won't start.

You can start an image in two places: the box on Home, or the Image app. Both take a plain-language prompt. You don't need a perfect one. Write what you want in a sentence or two, then refine from the results. This is text to image: you can write in any language, and the image can include text.

## Create a new image or edit an existing one?

Ideogram's Image app supports text-to-image generation and, on models that accept references, image editing from a prompt. Choose the starting point that matches your task:

| Your task                                         | Where to start                                                                                                  |
| ------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- |
| Make a new picture, poster or illustration        | Enter a text prompt on Home or in the Image app.                                                                |
| Change a photo or combine reference images        | Add [reference images](/create/reference-images.md), choose a model that accepts them, and describe the change. |
| Change a selected part of a picture               | Use [Masks and Inpainting](/edit/masks-and-inpainting.md) to mark the area.                                     |
| Add editable text or arrange several image layers | Open [Studio](/edit/studio.md).                                                                                 |
| Make a logo or transparent cutout                 | Follow [Make a Logo](/create/logos.md) or [Transparent Backgrounds](/create/transparent-backgrounds.md).        |

## Start from Home

The box under **What will you design?** is the fastest way in.

1. Type what you want to see. To use the example prompt that cycles in the box, press **Tab**.
2. Pick **Image** in the type picker.
3. Optional: select **Add reference images** and attach up to 9 images. Files must be JPEG, PNG or WebP, up to 25 MB each.
4. Press **Enter** or select the arrow button. On mobile, Enter adds a new line, so tap the arrow.

To make a clip instead, pick **Video** in step 2 (see [Create Videos](/create/video.md)). You can also paste or drag reference images into the box.

Ideogram opens the Image app on **Auto**, copies in your prompt and images, and starts the run. You don't need to click again. Auto picks a model for you; see [Models](/create/models.md#auto).

## Open the Image app

Select **Image** in the sidebar. The app opens on **Auto**. On desktop, the settings panel sits beside the **Assets** feed. On a phone, expand the **Create** sheet at the bottom to enter inputs, then collapse it to see your results.

### Generate an image in the Image app

1. Select **Image** in the sidebar.
2. Optional: open the **Model** picker and choose a model. The app starts on **Auto**, which picks one for you. See [Pick a model](#pick-a-model).
3. Type what you want to see in the **Prompt** box. See [Write a prompt](#write-a-prompt).
4. Optional: change the settings below the prompt. See [Choose settings](#choose-settings).
5. Select **Generate**. The cost in credits shows on the button.

   <figure><img src="https://1799634369-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzjhNby3LLsIikYuvxAJP%2Fuploads%2Fgit-blob-db65e41b24a4a985e55351b237ca2481b3f7fc7b%2Fcreate-images-1.png?alt=media" alt="Image app on Auto with a prompt filled in, four results in My assets, and the Generate button showing about 24 credits highlighted"><figcaption><p>Write a prompt, then select <strong>Generate</strong>. The cost shows on the button. In the current app, the results feed is labeled <strong>Assets</strong>.</p></figcaption></figure>
6. Find your images in the **Assets** feed. See [Find your results](#find-your-results).

In the Image app, press **Cmd+Enter** on Mac or **Ctrl+Enter** on Windows to run a generation. Plain Enter adds a line to your prompt. Home uses a different shortcut: Enter submits on desktop, while the arrow submits on mobile.

## Pick a model

Open the **Model** picker at the top of the panel. It lists:

* **Auto**, always first
* **Recent**: models you used lately
* **Team models** and **My models**: [custom models](/create/custom-models.md) you or your team trained
* **Pinned**: models you saved for quick access
* **All models**: Ideogram and partner models available to your account and workspace

Type in **Search all models...** to find one by name, or filter with the chips: **My models**, **Team models**, then one chip per provider. **Browse community models** at the bottom opens models other people have shared.

Each model has its own set of settings, and the panel changes to match when you switch. [Models](/create/models.md) compares them.

## Write a prompt

Type in the **Prompt** box ("What do you want to see?"). Describe the subject, the setting and the style. Put any words you want in the image in quotes, like `a poster that says "Summer Sale"`.

A few helpers sit in or near the prompt box:

* **Describe** turns an image into a prompt. See [Describe](/create/describe.md).
* Type `@` to pull one of your past images in as a reference, on models that take references. See [Reference Images](/create/reference-images.md).
* **Magic Prompt**, on Ideogram models, expands a short prompt into a detailed one. See [Magic Prompt](/create/image-settings.md#magic-prompt).

For more on writing prompts, read the [Prompting Guide](/prompting/prompting.md) and [Text in Images](/prompting/text-in-images.md).

## Choose settings

Below the prompt you'll find the settings for the model you picked: aspect ratio, number of images, quality and more. Extra options such as **Seed** sit under **Advanced settings**, a collapsed row near the bottom of the panel. The defaults work well, so you can skip this step. [Image Settings](/create/image-settings.md) explains each one.

## Generate

Select **Generate**. The button shows the cost in credits on its right side. On Auto it shows an estimate with a `~` in front, because Ideogram picks the model when the run starts. Hover over the number to see whether it's an estimate or the exact cost. To learn how credits work, see [Available Plans](/plans-and-billing/available-plans.md#how-credits-work).

The button reads "Starting…" while Ideogram sends the request. On a free plan with no bought credits, picking a premium model opens an upgrade window instead. So does running out of free credits.

### Why won't Generate start?

| What you see                                     | What to check                                                                                                                                 |
| ------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------- |
| **Enter a prompt before generating.**            | Add a description; a blank prompt or spaces alone won't run.                                                                                  |
| **Choose a model before generating.**            | Select the custom model you want to use.                                                                                                      |
| **A mentioned image is not used by this model.** | Remove its `@` mention or switch to a model that accepts that reference. See [Reference Images](/create/reference-images.md).                 |
| An upgrade window                                | Check whether the app needs premium access or your free credits have run out. See [Plans and Credits](/plans-and-billing/available-plans.md). |
| No credit amount on the button                   | The app doesn't yet have a ready quote. Complete the required inputs and check again; an absent amount doesn't mean the run is free.          |

Keep the inputs visible while you resolve an error. **Starting…** means the request is being sent; it is different from a completed generation.

## Find your results

Results appear in the app's **Assets** feed. On desktop this is beside the inputs; on mobile it sits behind the collapsible Create sheet. A "Generation started" notice appears with **Jump to new generation**, and blank tiles hold each image's place until it's done. If a run fails, you'll see "Could not generate that image. Please try again."

Switch to **List** layout to see each run's available actions:

* **Copy prompt** and **Reuse prompt**
* **Recreate**, which restores the supported inputs and settings from that run to its app
* **Download**, which saves every image in the run

Click an image to open it. From there you can **Edit**, **Remove BG**, **Upscale**, **Use as source**, **Recreate**, **Open in app**, **Add to collection**, **Like**, **Share** or **Download**.

Use the toolbar above your results to search ("Search your generations") or filter to the current app. Open **Layout** to choose **List**, **Grid** or **Mosaic**, and use **Size** to adjust the tiles.

{% hint style="info" %}
On Plus, Pro, Team and Enterprise, new images stay unpublished until you publish them to the community. Free and legacy Basic generations are public. Sharing a collection can separately grant access to other people; see [Share and Publish](/organize-and-share/share-and-publish.md). To publish an image:

1. Open the image.
2. Select **More options**.
3. Select **Manage**.
4. Select **Publish to community**.
   {% endhint %}

On desktop, the **API** tab next to **Assets** shows a supported request example where available. Use the [API guide](/plans-and-billing/ideogram-api.md) for key setup and the public developer reference for request details.

### Copy prompt, Reuse prompt or Recreate?

Use **Copy prompt** to put the text on your clipboard. In the List feed, **Reuse prompt** opens the app with the prompt and its available source references; it doesn't restore every setting. Use **Recreate** when you want the original app, supported settings and source inputs loaded for another run; review them before generating again.

Recreate is available for supported generated assets whose original app is still accessible. An uploaded file or a result from an unavailable app may not offer it. Recreating a run prepares a new request; it doesn't restore a deleted file or guarantee the same output.

To find older work outside the app, open **Library** → **Generations**. Uploaded source files are under **Library** → **Uploads**. See [Library and Favorites](/organize-and-share/assets.md).

## Next steps

* Change part of an image: open it and select **Edit**.
* Save it: see [Download](/organize-and-share/download.md) for file types and sizes.
* Need a clear background? See [Transparent Backgrounds](/create/transparent-backgrounds.md).

## In this topic: creating images

1. **Create Images** (this page): make your first image
2. [Models](/create/models.md): pick the model for the job
3. [Image Settings](/create/image-settings.md): set size, quality and other options
4. [Prompting Basics](/prompting/prompting.md): write a prompt that works
5. [Fix Common Problems](/prompting/fix-common-problems.md): fix results that come out wrong