# Image Generation

Generate pictures of your companion in the scenes you describe. Your **prompt** is the text that tells the image generator what to create.

## **Three ways to start**

You can ask for a photo right in the chat, or open the full generator for more control.

### **The camera icon in Messages view**

Tap the **camera icon** in the message bar, then choose **Image**.

![Camera menu on the Image tab with the Model selector (V3 2K) and the camera icon both outlined in coral red](https://cdn.sanity.io/images/bf00glst/production/b70f255a353e5040a9950bed2d56e38a1c56a115-730x938.png)

| **Option** | **What it does** |
| --- | --- |
| **Send me a photo** | Requests a photo based on the current scene. |
| **Send me a photo of...** | Lets you describe the photo you want. Finish the sentence, then send. |
| **Generate photos** | Opens the full generator for more control. |

For your first photo, try **Send me a photo of...** with:

> You sitting at a café table, wearing a blue sweater, with morning light through the window.

In a group chat, name the companion you want:

> Sophia: send me a photo of you at the bar.

Photos with more than one companion require the **V3 Model**.

### **The Generate button in messages view**

Open the chat panel, go to the **Gallery** tab, and tap **Generate**. This opens the full generator directly.

![Chat sidebar on the gallery tab, with the background blurred, showing the red Generate button](https://cdn.sanity.io/images/bf00glst/production/b04ca3898d6ce9b3c3ea06022c66202f0b64e27e-627x1436.png)

### The Create page

Tap **Create** in the bottom bar (nectar.ai/create). Your past images are there too.

![Left navigation menu with the Generate item highlighted among Discover, Chats, Create, Community and Upgrade](https://cdn.sanity.io/images/bf00glst/production/19c56e676423de1abde82ec0cbcaeb57eb005741-466x580.png)

## **Using the generator**

The full generator lets you choose the companions, scene, and image settings. For your first companion image, choose **Nectar Companion Model** and one companion.

> 💡

> Some fields change with your selected model.

### **1. Type**

Choose **Image**.

### **2. Model**

**Model** selects the image model you want to use. Choose **Nectar Companion Model** to generate an image of a selected companion.

Other options include **Nectar Base Model** (SFW, the default), **Nectar Anime Model**, **Nectar Realistic Model**, **Nectar 2.5D Model**, **Nectar Experimental Realistic Model**, and **Nectar Ultimate Asian, Anime and Realistic**.

If you need beta access, submit a support ticket in our Discord.

Image models are separate from the chat models that write replies. “Ultimate” in an image model’s name is not your subscription plan.

Your choice also affects how you write the prompt. See **[Writing better prompts](/media-generation/image-generation#writing-better-prompts)** below.

### **3. Companions**

Choose who appears in the picture. You can select up to **10 companions** for a V3 image.

Check the selection even when you open the generator from a chat.

1. Tap the **Companions** field.

1. Use **Discover** to find public companions or **Saved** to find your bookmarks and likes.

1. Select the companions you want.

1. Remove anyone with the **X** on their chip.

1. Tap **Accept**.

![Companion picker with search, For You filter and NSFW toggle, one companion selected and an Accept (1/10) button](https://cdn.sanity.io/images/bf00glst/production/f25aac747726b93ad34b94a63773f35f382c399c-996x1424.png)

Start with one companion while you learn the generator. For group images, see [Multiple companions](/media-generation/writing-image-prompts#multiple-companions) on Writing Image Prompts.

### 4. Version

After choosing a companion, pick a version such as V3 in the **Model** picker.

These are versions of the companion image model, not separate visual styles.

**V3** and **V3 2K** give you the most detail and the most control over the scene. Use one of these for the steps on this page.

Your plan also includes an image allowance. Your account shows which versions it covers and the price of each paid generation.

Pick a shape under **Dimensions**: **1:1** square, **2:3** portrait or **3:2** landscape.

### **5. Prompt**

Describe the companion’s clothes, action, surroundings, and lighting by typing in the text box.

```
Mika sits at a café table in a blue sweater, holding a white coffee mug. Morning light comes through the window. Waist-up view.
```

> 💡

> **[Enhance](/media-generation/customizations)** expands a rough idea into a full prompt. Check its displayed price before using it. Read the result before tapping **Create**.

On the **Video** type, suggestion chips sit under the prompt box in **Body** and **Scene** groups. Tap **Show more** for more.

Tap a chip to add it. Tap **Generate new ideas** for a fresh set.

![Create page on the Video type with prompt suggestion chips and Generate new ideas outlined in coral red](https://cdn.sanity.io/images/bf00glst/production/79bf64cc5b0e8fc8eab99918a15a90f08a6aced3-370x545.png)

### **6. Icks**

**Icks** describes what you want left out of the image. This is also called a negative prompt.

For example:

```
text, watermark
```

Keep this consistent with your main prompt. If you ask for a red coat, do not also exclude red.

The field appears on supported options, including the **Realistic** and **Companion V2 models**. It does not appear on the **Companion V3 model**.

### **7. Customizations**

**Customizations** are ready-made tiles for faces, outfits, poses, and effects. Available categories depend on the model.

Choose up to **three per image**. Keep your selections consistent with your prompt.

See [Customizations](/media-generation/customizations).

![Customizations panel on the Outfits (10) tab showing outfit cards tagged Ult or Pro, next to Faces, Poses and Effects tabs](https://cdn.sanity.io/images/bf00glst/production/6ba70fa5dadf23670584b29dcdb2df47632b4b49-1536x567.png)

### **8. Create**

Review your companions, prompt, and settings. Check the displayed price before a paid generation, then tap **Create**.

To refine the result, change one detail and generate again.

---

## **Writing better prompts**

Your prompt is the biggest thing you control. Start with your companion, their clothes, one action, the place, and the light. Change one detail at a time so you can tell what helped.

For the prompt format each model likes, weights, V3 prompts, pictures with several companions, and quick fixes, see [Writing Image Prompts](/media-generation/writing-image-prompts).

### **Generating in the background**

Images generate in the background while you chat and browse.

- **Pro:** Up to two images at once.

- **Ultimate:** Up to four images at once.

---

## **What you can do with a generated image**

- **Turn it into a video.** Use the video button on the image. See [Video Generation](/media-generation/video-generation).

- **Make it the profile picture.** Open the image and choose **Submit as profile picture**.

- **Add it to the gallery.** Review the destination and visibility options before featuring it.

- **Let it set the mood.** Your chat background changes to your most recent generated image.

- **Attach it to a greeting.** Use **Greeting Media** in the companion editor.

- **Save it to a folder.** Organize images into collections you can return to from **Saved Items**.

- **Download it.** Open the creation from **My Creations** on the **Create** page, tap the three dots, then **Download**.

Next: [Customizations](/media-generation/customizations) · [Writing Image Prompts](/media-generation/writing-image-prompts) · Stuck? [Troubleshooting and Getting Help](/help-center/troubleshooting-and-getting-help)