Hulab

AI photo & video studio

Consistent photos and videos of one character, generated in batches with the best AI models.

Add her face and body once. Drop in the scenes you want. Hulab reads every reference, writes the prompt and puts the same woman into each shot.

8
reference photos per character
up to 5 face + 3 body
5
photo and video models
in one workspace
4K
maximum photo resolution
1K, 2K or 4K per row
16
generations in parallel
per batch

How it works

From a few photos to a whole set

  1. 1

    Add her references

    Up to five face photos from different angles and up to three full-body shots. They define who she is.

  2. 2

    Drop in scenes

    Any photos you like: a mirror selfie, a café, a beach. Each one becomes a row in the batch.

  3. scene_type mirror selfie
    pose leaning on the sink
    light soft window light
    outfit navy knit dress
    3

    Hulab writes the prompts

    Gemini reads pose, outfit, framing and light. Hulab builds a prompt that keeps her identity locked.

  4. 4

    Generate the batch

    Pick a model and resolution, see the cost, run. Retry any row on its own.

Under the hood

Every image gets a job

Gemini breaks each scene down into pose, expression, outfit, light and framing — and never describes her appearance. Hulab then numbers every reference and tells the model exactly what to take from it.

  • Her face and hair come from the face photos only
  • Body proportions come from the body photos
  • The person in the scene is a stand-in and is fully replaced
Prompt · row 3assembled
@Image1 — her face. Identity only.
@Image2 — her body: a plain full-body reference. Take only the shape of her body from it.
@Image3 — the scene to recreate. The woman in it is a stand-in.

A new photograph of the woman shown in @Image1. Her build and body proportions follow @Image2…

This is a mirror selfie. She leans on the sink…

Features

Built for producing at volume

Batch runs

Dozens of scenes in one go, up to sixteen at a time. Every row can be rerun or tweaked on its own.

Identity lock

Prompts tie face, hair and skin tone to your references only and tell the model to leave the scene's stand-in, tattoos included, behind.

Two scene modes

Scene copy reproduces a reference closely. Description rebuilds the scene from text alone.

Model choice

Nano Banana Pro, Seedream, Wan and more. Switch per batch; prompts adapt to each model's syntax.

Cost before you run

Every batch and video shows its credit cost before you confirm, with a spending cap on top.

Your library

Results, assets, prompt library and full history in one place. Export a whole batch as a ZIP.

Video

Put her into motion

Character swap

Keep a clip exactly as it is and replace the woman in it with yours.

Motion transfer

Movement and camera come from a reference video; she and her look come from your photos.

Bring a photo to life

Turn any result into a short clip.

Pricing

Credit packages

Starter

$10
100 credits
$0.100 per credit

Studio

$160
2,000 credits
$0.080 per credit

Prices in USD. Every generation shows its credit cost before you confirm. See the Terms.

FAQ

Questions

How do credits work?

Hulab is prepaid. You buy a credit package; every generation shows how many credits it will use before you confirm it.

How do I pay?

In cryptocurrency, through a third-party payment processor. Credits appear on your balance once the payment is confirmed on-chain.

What if a generation fails?

If the AI provider fails and produces no result, the credits are returned to your balance automatically.

Is my content used to train AI?

No. Your uploads and results are used only to run the service for you and are never sold or used for training.

Who can use Hulab?

Adults only — you must be 18 or older. Upload only people who have given their consent.

One character. Every scene.