Daz Studio can now turn a scene into an AI image or video

Daz 3D has released the AI Scene Enhancer, a tool for generating images and video built into Daz Studio. It uses the scene assembled in the program as its basis: characters, poses, camera and lighting. Cloud models from OpenAI and xAI repaint it in a chosen style or turn it into a short clip. The source can be the viewport, a finished render or any image on disk, so there is no need to save a picture and upload it to a third-party generator by hand.

The tool became available for download on 10 November at 10:00 MST in Daz Install Manager (DIM) for Windows and Mac. Installation is free, while each generation deducts credits, from 2 to 200 and more at a time. Generation happens only in the cloud, so an internet connection and a sign-in to a Daz account are required.

The scene as a basis for generation

In Daz Studio the scene is assembled before the prompt is written. The artist places characters, builds poses, frames the shot and sets the light, and the model receives all of this as a visual basis that helps guide the result. You can work straight from the viewport while assembling the scene, with a final render or with any other image on your computer.

The tool is located in the Render > AI Scene Enhancer menu. Iray stays in place and the usual rendering works as before. According to Daz, this is a separate tool inside Daz Studio and should not be confused with Daz AI Studio.

The window is split into three tabs. Generation happens in AI Scene Enhancer, clips are assembled into a sequence in Video Stitcher, and they are cut down in Trimmer.

20 styles and 5 effects at launch

The ready-made styles cover photorealism, handcrafted textures, illustration, stylised 3D and animation. A few examples:

• photographic: Moody Film Noir, Infrared Film Photography, Wet Plate Photography, Dreamy Experimental Double Exposure;
• textural: Felt & Yarn Diorama (a diorama of felt and yarn) and Cross Stitch Textile Art (cross-stitch embroidery);
• illustrative: Graphic Novel and Watercolor Storybook;
• stylised 3D: Stylized 3D Animated Feature in the spirit of a feature-length animated film;
• animation: Context-Aware Shonen Anime Choreography and Gentle Breeze Animation.

Five post-processing effects provide additional ways to refine an image. Adaptive Typography and Layered Poster Design, judging by the name, lays out the frame as a poster with typography. The other four add anamorphic flares, rain, snow and volumetric light with rays. Daz promises to expand the set of styles and effects over time.

Presets can be used as a starting point, or you can describe the look you want in words. In the Manage Styles section your own prompts are saved, edited and applied to other images. This comes in handy when the same style has to be maintained across a series of frames.

What can be adjusted

Images are generated by OpenAI and xAI at two resolutions: Standard 1K and High 2K. OpenAI offers three variation modes: Standard, More Creative and More Like Original. Judging by the names, the second lets the model stray further from the source, while the third keeps the result closer to the scene. OpenAI also offers a transparent background, useful for example for a character that will later be placed into a composition. The Auto Save function puts results into a chosen folder by itself, and it can be switched on or off.

Video is generated by xAI. You can choose duration, aspect ratio (matching the source, 16:9, 1:1, 9:16 and others) and resolution: 480p Draft, 720p Standard or 1080p High. Draft quality logically serves for finding the motion, with the final one run on the selected variant. There is audio generation and five ready-made voices for voice-over, and you can add your own.

Best Quality mode works in two steps. First OpenAI creates an image from a style or prompt, then Grok from xAI animates it. This mode consumes more credits.

Video Stitcher imports clips from the computer, adds transitions, background music, titles and captions, and exports the result at 24, 30 or 60 frames per second. A short clip made of several generations can be assembled without leaving Daz Studio.

Credits, rights and data

New users receive starter credits for free, while existing ones get a one-off allocation. Free credits are valid for 30 days. After that a monthly subscription or prepaid packs are required, and subscription credits are valid for one month. Daz publishes prices on the generative credits page. How many images or seconds of video a pack yields in practice cannot be worked out from the announcement. Prompts blocked by moderation do not spend credits, either for images or for video. Failed image generations are also free. A video request already sent to the provider, however, may deduct credits even if the generation does not complete.

Rights to the result belong to the user, and images and video can be used commercially. The company separately addresses the question of renders of purchased content:

«The Daz license agreement allows licensed Daz content to be passed to an AI in the form of a render».

So states the Daz 3D announcement, citing the Daz EULA. According to the company, users' images and renders are not used to train AI, and Daz itself does not store them. The source image and the prompt do go to the chosen provider, though, and its storage rules may differ.

What has changed compared with Daz AI Studio

In April 2024 Daz launched Daz AI Studio together with Stability AI. It was a separate generator built on the SDXL 1.0 model that made pictures from a text description in the style of Daz renders. Some of the parameters, including lighting and environment, were moved there from the prompt into a graphical interface. The generator was aimed at those who lack the skills and hardware for 3D rendering. AI Scene Enhancer works inside Daz Studio and is aimed at those who have already assembled a scene and hand the generation over to external models from OpenAI and xAI.

Demand for a tool like this had been voiced in the community before. In March 2025, on the Daz forum, a user proposed using a diffusion model as a render engine, with the 3D scene serving as the controlling input. In their view this would greatly raise the value of the store's older content. One participant replied that they already push quick Filament and OpenGL renders through Stable Diffusion in image-to-image mode by hand. Another asked whether this meant passing the current viewport view to a generator without saving and importing the picture. Now such a hand-off has become a menu item.

Who will find it useful and what the limitations are

The biggest winners are likely to be illustrators and comic authors working in Daz, who were already refining their renders in third-party generators: files no longer have to be moved between programs. A scene can be quickly checked in several styles, from noir to embroidery, and an animatic with voice-over can be put together. The tool may also be of interest to owners of weak graphics cards: a powerful card is not needed for the generation itself, although the computer still has to handle Daz Studio and local rendering. Those who value Iray's physically correct light and predictable results have no reason to switch to AI processing.

There are several limitations. Working offline is not possible. Costs depend on the number of attempts; Best Quality is officially more expensive than ordinary generation, and judging by the spread from 2 to over 200 credits, video overall will also cost more than images. The set of models and the data storage rules are determined by the external providers. How accurately the generation preserves faces and the recognisability of characters from frame to frame is not disclosed by Daz in the announcement: the company says only that the scene helps guide the result. For serial work with a single character it makes sense to test this on the starter credits before they expire after 30 days.