Create images, videos, and custom characters in Venice AI

What Venice gives you and where to start

Venice is built to let you create and explore without the limits other AI providers impose. Everything starts in the chat window, where you can see every model available to you: text and chat models, image generation, and video generation that can be prompted from text or from an existing image.

Venice also offers uncensored options. Venice Uncensored 1.1 is the most uncensored text model, uncensored image generation models are available to pro users, and the open source video models do not censor either. If you just want to talk, choose the auto setting and start typing.

Generate a prompt and create your first image

A fast way to get a strong image is to let the AI write the prompt for you. Ask the chat model for an image prompt, then feed that prompt into an image model.

Generate an image prompt for a character that absolutely loves pizza, pepperoni in particular.

An AI-written prompt tends to be more detailed than one you'd write by hand, catching specifics you might skip. Paste it into the image generator (Hidream is the powerful default) and it produces four variants. In settings you can change the number of variants, the aspect ratio, and apply a style, which starts at none.

Turn an image into video and compare models

Open an image you like and click create video. The default prompt is "bring this image to life," with the image-to-video model 1.2.5 selected by default. Replace it with your own direction, or click the magic wand to have AI enhance the prompt with extra detail and adjectives.

This pepperoni pizza lover gives us a tour of his pizzeria and the highest quality pizza in town.

The fork button lets you split a conversation onto two paths, which is handy for running the same idea through different models. In the video, one branch used model 1.2.5 and another used Veo 3.1 fast, so you can compare results side by side. Each generation shows its credit cost up front, for example 83 credits for one of these video clips.

Extend a clip with continuations

Once you have a video you like, you can keep the shot going. Copy the last frame and continue, then describe what happens next.

Man chokes on the pizza because it's too hot. The people behind him yell and freak out and come to save him.

Before sending, check the same settings you control elsewhere: you can switch to a vertical orientation and add more variants, just like with image generation.

Shape AI behavior with system prompts

Open the settings icon to add a system prompt, which works like telling a character how to act. A system prompt can define a marketing strategist, an assistant that turns long transcriptions into outlines, or anything else. Containing the conversation this way produces more focused results.

From the same panel you can add a voice for spoken responses and adjust temperature and top P for more creative output, with the caveat that higher creativity can lead to hallucinated answers. Turn on a "pizza lover" system prompt and the assistant argues passionately for pizza over pasta, behavior it only exhibits because the prompt told it to.

Save a reusable character

The characters menu holds a public list of characters specialized in different subjects. You can also build your own. Create a character, give it a short description like "this man absolutely loves pepperoni pizza," and auto-generate a personality, image, and instructions.

You can add an intro statement and reference material, such as a restaurant menu so the character knows exactly what is served, then save and start chatting. To refine it later, edit the character and adjust the instructions to set a specific attitude or demeanor.

Assemble a complete video ad and manage credits

To build a finished ad, start by asking a text model for a video generation prompt, for example a sales ad for pizza, condensed into paragraph format. Generate an image of the opening shot, click create video, paste in the prompt, and pick a model such as Veo 3.1 fast.

That final ad cost 132 credits, roughly $1.32, which gives you a sense of real pricing. Click the credits button any time to see your balance. You can buy more with card or crypto, or stake the VVV token from the token dashboard to receive refreshing daily Venice credits. To share what you make and see what's coming, join the Venice community on Discord.

Key takeaways

• Let an AI write your image prompts for more detail, then chain image to video to extend clips from the last frame.
• Forking a conversation lets you run the same idea through different models and compare results.
• System prompts control an AI's personality and keep conversations focused; save a good one as a reusable character.
• Every generation shows its credit cost up front, and you can refill with card, crypto, or by staking for daily credits.