Question 1

Changelog

Accepted Answer

2026-09-08

Video

Videos in the Videos page now load and play more smoothly, for a better browsing experience.

Model builder You can now include a brief when creating a custom model, making it easier to give context and guidance right from the start.

2026-09-03

Global

We've simplified how credits work on Seelab. Image and video credits are now unified into a single credit pool. What's changing:

Your video credits have been automatically transferred into your image credits — no action needed on your end.
One credit balance now covers both image and video generation.
Your total available credits remain the same; only the way they're organized has changed.

2026-09-02

Model builder

Creating a custom model is now faster! It takes just a few seconds to a few minutes.

2026-09-01

Downloads

Multi-file streaming downloads are now available across the entire app, for faster and more reliable exports.

Composer

Layers can now go up to 4096px (previously limited to a smaller size).

Home Page

Fixed a bug where the link to the video tool wasn't redirecting to the correct dedicated page.

2026-08-26

Video

Generate from a custom model You can now generate videos directly from your custom models, for creations that stay true to your product identity or characters.

Editor

Deleting generation history Fixed a bug that prevented deleting items from the generation history in

Question 2

What is a Seelab model?

Accepted Answer

A Seelab model learns your product, your character or your style from a few images. Every visual you then generate with it stays true to the original: same shape, same label, same face, same look and feel. This is what lets you produce hundreds of brand visuals without a photo shoot.

![The Custom models space](https://storage.crisp.chat/users/helpdesk/website/-/6/8/0/3/6803ea00be5f5800/image_1y4o53b.jpg =1000xauto)

Three types of models

![The three types of models](https://storage.crisp.chat/users/helpdesk/website/-/6/8/0/3/6803ea00be5f5800/image_3k90dw.jpg =600xauto)

Portraits & Avatars Recognizable, consistent characters: mascots, virtual models, brand ambassadors.

Products & Objects Your products, faithful visual after visual, for e-commerce, social media and campaigns.

Styles & Photography Your brand's visual DNA, in photography and illustration alike: palette, lighting, composition.

What you can do with a Seelab model

Generate images: choose your model in the model selector and describe the scene.
Generate videos: one training creates both your image model and your video model.
Insert your product into an existing image, with Replace in the editor.
Upscale an image with precision, by selecting your model in Magnify.
Combine your model with reference images: the product comes from the model, the mood from the reference.

How it works, in short

You choose the model type.
You import a few images of your

Question 3

Using up to 10 reference images

Accepted Answer

Reference images let you show Seelab what words describe poorly: a specific product, a setting, a lighting, a style. You can add up to 10 per generation and combine them: three products and a setting, a character and an outfit, your product and the mood of a campaign.

![The Reference Images panel](https://storage.crisp.chat/users/helpdesk/website/-/6/8/0/3/6803ea00be5f5800/02-07reference-images-panelen_1m2n10.jpg =1000xauto)

What references let you do

Bring several products together in one scene: add your products and a setting, and Seelab brings them together in a single image, lighting included.
Place an object in a setting: import an object, then describe where and how to place it ("the perfume in a modern bathroom").
Apply a style or a mood: one image provides the lighting, palette, grain or composition to reproduce on another subject.
Keep a character consistent: the same face or object in different poses, framings or outfits.
Dress or equip a subject: "Apply the shirt to the woman"; a garment, an accessory or a pack worn or held by a model.
Iterate on an existing image: change the setting, the framing or the season of an image you have already generated.

Adding reference images

Click the image button, left of the prompt. The Reference Images panel opens.
Drag in your images, or click to open the Resource Hub: your imports, your Seelab model datasets, your collections and your latest generations.
**Add

Question 4

Understanding credits

Accepted Answer

Every generation or edit uses credits. The cost depends on the tool, the model and the settings, and it is always shown before you launch: you know what you are spending before you click.

Where to see the cost before launching

Creation: above the green generate button.
Video: next to the Generate button.
Workflow: on the Run all or Run section button, updated with every change.

Tracking your usage

The Credits gauge, at the top of the screen, tracks your usage.
The Credits tab in Organization settings, in your avatar menu, gives the details for the organization.

What affects the cost

For images: the model, the resolution (1k, 2k or 4k), the quality when the model offers it, the number of images, and reference images, which slightly increase the cost.

For video: the engine, the resolution, the duration and the sound. The cost is calculated per second of video.

| Image model | Credits per image | | --- | --- | | GPT Image 2.5 Flare | 1 to 18 | | Nano Banana 2 Lite, Nano Banana | 2 | | SeeDream 5.0 Lite, SeeDream 4.5 | 2 to 4 | | Nano Banana 2.1 | 2.3 to 7.7 | | Flux 2 PRO | 2 to 8 | | SeeDream 5.0 Pro | 4 to 8 | | Nano Banana Pro | 8 to 21 | | GPT Image 2.5 Sunburst | 2 to 36 |

Details for each model, image and video: Image models in detail and Video models in detail.

Spend

Question 5

Generate your first image

Accepted Answer

Describe the scene you want, choose a model and click the green button: your images appear at the top of your feed in a few seconds. This article walks you through your first generation, step by step.

![The Seelab Creation screen](https://storage.crisp.chat/users/helpdesk/website/-/6/8/0/3/6803ea00be5f5800/02-01creationoverviewen_1lyxij6.jpg =1000xauto)

The Creation screen at a glance

The navigation bar: Creation, Collections, Edition, Video, Custom models and Workflow.
The project, credits and account, at the top right. The selector shows the active project. The Credits gauge tracks your usage. Help opens the help center.
The prompt bar: reference images on the left, your text in the center (up to 3,000 characters).
The model, the palette and the settings, then the green button to generate. The credit cost is shown just above it.
The history of your generations: search, filters by date and model, list or grid view.
The Collections panel, on the side, to file your images and create collections.

Generate an image in five steps

Write your prompt. Describe the scene, the lighting, the framing and how the visual will be used. Type @ to insert a model, an image or a color without leaving the keyboard.
Choose the model. Click the model name below the prompt. Your Seelab models are listed first.
Add reference images if needed. The image button, left of the prompt, accepts up to 10 images. See [Using up to

Question 6

Which image model to choose?

Accepted Answer

Start from your need, not from the model name. For a visual that stays true to your product, use your Seelab model; for everything else, each market model has its own strength.

The right model for each need

| Your need | The model | Why | | --- | --- | --- | | A visual faithful to your product, character or style | Your Seelab model | Respects shape, logo and materials | | A staged scene or a retouch by instruction | Nano Banana 2.1 | Best quality / credits ratio | | A complex hero visual, no compromise | Nano Banana Pro | Maximum precision | | Legible text: poster, packaging, tagline | GPT Image 2.5 | Most reliable typography | | A realistic photo, an editorial mood | SeeDream 5.0 Pro | Natural light and skin | | Consistent variations in a workflow | Flux 2 PRO | Stable and repeatable | | Exploring ideas very quickly | Nano Banana 2 Lite or SeeDream 5.0 Lite | 2 credits per image |

Choosing the model

![The model selector](https://storage.crisp.chat/users/helpdesk/website/-/6/8/0/3/6803ea00be5f5800/02-02market-model-picker_co59zu.jpg =1000xauto)

Click the model name, in the prompt bar.
Choose: your Seelab models are listed first, then the market models.
Check the available settings: formats, resolutions and quality vary from one model to another.

| With your Seelab model, describe the scene, never the product: the model already knows it.

A two-step method

Explore with a fast, economical model: run sever

Featured & Frequently Read Articles

Browse All Categories