Uncensored AI Image Generator:
Unleash Your Creativity

Tired of running into wall when you try to generate a photo in ChatGPT? You make prompts but get refusals? That can be really frustrating when you face political, religious, and other kinds of restrictions.

Image Generators in Secret Chat

Below you can see the picture generators that our app has. They are classified by the level of their permissions. Green are the most uncensored, yellow are moderately unfiltered, red have the most restrictions.

• 🟢Venice LST
• 🟢Venice Anime
• 🟢Venice Qwen Image
• 🟡Venice Stable Diffusion 3.5
• 🟡Grok 2 Image
• 🟡Nano Banana
• 🔴Dall-E3
• 🔴GPT-5

An Example of a Photo Generation

Let’s make an experiment. I ask Grok 2 Image a very simple prompt that can demonstrate how much unfiltered our models are, ‘a girl in only bikini full-height, standing with her back to us’.

GPT-5 and Dall-E3 refused to generate an image by this prompt.

We get the following result:

Grok also kindly provided us with a refined prompt: ‘A high-resolution photograph of a young woman standing on a beach at sunset, wearing a vibrant red bikini. She is facing away from the viewer, with her back to us, and her long, wavy brown hair falls down her back. The background features a calm ocean and a warm, setting sun, casting a golden light across the scene. The beach is relatively empty, with no distracting elements, focusing solely on the girl and the natural surroundings. The atmosphere is serene and tranquil, with the ocean gently visible in the background without any foreground elements that might distract from the main subject.’

I continued with this detailed prompt. GPT-5 and Dall-E3 refused again. Here’s another version by Grok:

Nano Banana:

Venice LST:

Venice Stable Diffusion 3.5:

Venice Qwen Image:

AI Prompt Generator

While Grok 2 Image provides us with a detailed prompt, other models don’t. That’s why we’ve added a handy prompt generator for other models.

Image Settings for Venice

For Venice’s models we provide convenient settings.

Aspect ratio

• Square (default)
• Landscape (3:2)
• Cinema (16:9)
• Tall (9:16)
• Portrait (2:3)
• Instagram (4:5)

Adherence

This is a CFG scale of how strictly the image generator will stick to prompt. From “Creative” to “Strict”.

Variants quantity

How many images do you want to generate at a time? Default 1, maximum 4.

Hide watermark

You can leave or hide watermark, as you wish.

Image style

Default is no stylizing, just photo, and there are 70+ image styles. For example, “3D Model”, “Analog Film”, “Cinematic”, “Comic Book”, “Cubist”, “Fantasy Art”, “GTA”, “Impressionist”, “Long Exposure”, “Minecraft”, “Pop Art”, “Watercolor”.

Examples of image styles (only Venice)

Let’s use a prompt “A fluffy white rabbit with its paws firmly gripping the handlebars and feet on the pedals of a vintage red bicycle. The bicycle is well-defined with a chain and spokes. The rabbit is riding along a beautiful, winding forest path with soft lighting”. I will use Venice Stable Diffusion 3.5 (SD) and Venice Qwen Image (Q)

Image Edit

Some of the models allow to upload and edit images.

Nano Banana

The most capable and creative when it goes to images edit. Moderate restrictions.

Let's generate our initial image. “A model girl standing in full height”

Then I ask “Make this girl sit down”

Then I ask “Make this girl wave her hand standing with her side to us”

I think, it’s perfect!

Venice Qwen Image

It has less restrictions than Nano Banana, but not so capable.

“Make this girl sit down”

“Make this girl wear umbrella”

GPT-5

The most restricted and not very capable.

“Make this girl sit down”

Frequently Asked Questions