# Your first generation

Make your first AI image or video in Sloosh in four steps: open Studio, describe what you want, check the chips and press Generate.

Last updated: 2026-10-08

**Key facts**

- Your first Sloosh generation takes four steps: open Studio's [Image](https://sloosh.ai/studio/image) or [Video](https://sloosh.ai/studio/video) page, describe what you want, check the chips under the prompt, and press Generate.
- Sloosh Studio's Generate button shows the credit cost before you spend anything, and your first Sloosh workspace gets 20 free starter credits once.
- The default model in Sloosh Studio is GPT Image 2.5 Sunburst for images and Seedance 2.5 for video.
- Sloosh Studio results land in My Generations, above the prompt, and everything your workspace makes is also kept in [Assets](https://sloosh.ai/docs/library.md).

Web page: https://sloosh.ai/docs/first-generation

## Make something

![Studio's prompt box with a prompt typed in, the chips under it and the Generate button showing its cost](https://sloosh.ai/docs/studio-composer.webp)

_Under the prompt, left to right: + adds references, then the model, Enhance (the wand), the model's own settings, how many to make, and Generate with the cost of the whole batch._

1. Open Studio: [Image](https://sloosh.ai/studio/image) for a picture, [Video](https://sloosh.ai/studio/video) for a clip. Sign in if asked.
2. Describe what you want to see. The [prompt guide](https://sloosh.ai/docs/prompt-guide.md) has examples.
3. Check the chips (the small buttons) under the prompt. One picks the model; the others are that model's settings, such as aspect ratio or resolution, and how many outputs to make.
4. Press Generate. The button shows the credit cost first. A new workspace starts with 20 free starter credits.
5. Your result lands in My Generations, above the prompt. Open it to download it, copy it or keep editing.

## Not sure what to write?

Open the Templates tab in Studio. Each card is something another person made, with the prompt and settings behind it. Press Recreate to load them into your prompt box, then change what you like. See [Templates](https://sloosh.ai/docs/templates.md).

## If the result isn't right

- Change one thing in the prompt and generate again. If you rewrite everything, you can't tell what helped.
- Try another model with the same prompt. Each one reads prompts a little differently.
- If it must show your real product or person, add a photo as a [reference](https://sloosh.ai/docs/references.md).

## What to try next

- Add a photo of your product as a [reference](https://sloosh.ai/docs/references.md) so the result shows the real thing.
- Want the same kind of result again and again? Rebuild it as a [space in Spacelab](https://sloosh.ai/docs/spacelab.md), so you or anyone can make it from a form.

## Frequently asked questions

### How do I make an AI product photo in Sloosh?

Open Studio's [Image](https://sloosh.ai/studio/image) page, add a photo of your product as a [reference](https://sloosh.ai/docs/references.md), and describe the scene: the surface, the light and where the photo will be used. Say what must not change, such as the colour, logo or shape, so the result shows your real product rather than a lookalike.

### Which model should I use for my first image in Sloosh?

Keep the model Sloosh Studio picks for you, GPT Image 2.5 Sunburst, for your first image. If the result isn't right, run the same prompt on a second model and compare, because each model reads prompts a little differently. [Models](https://sloosh.ai/docs/models.md) lists what each one is best for.

### Can I make an AI video as my first generation in Sloosh?

Yes. Open Studio's [Video](https://sloosh.ai/studio/video) page, describe the shot and press Generate; the default video model is Seedance 2.5. Describe what moves, the camera move and the pace, as [Video prompts](https://sloosh.ai/docs/video-prompts.md) shows.

### What should I do if my first Sloosh generation doesn't look right?

Change one thing in the prompt and generate again, so you can tell what helped. If that doesn't fix it, try another model with the same prompt, or add a photo as a [reference](https://sloosh.ai/docs/references.md) when the result must show a real product or person.