How to direct AI images and videos with iClone 3D guidance

Most text-to-image and text-to-video tools struggle when you ask for precise poses, camera moves, or consistent characters across shots. That’s where 3D guidance comes in. By using iClone together with AI Studio, you can lay out full 3D scenes, animate characters, and then let AI handle the final visual style.

What 3D-guided AI generation actually does

Instead of relying only on prompts, you first build and animate your scene in 3D. Characters, props, cameras, and even lighting are set up in iClone. Then AI Studio reads this 3D information as a strong guide and generates final images or videos in the style you choose.

This approach gives you:

• Precise control over character poses and motion
• Stable character identity across many shots
• Consistent backgrounds and lighting
• Cinematic camera moves that are hard to describe with text alone

Building scenes with characters, props, and poses

You start by assembling your scene in iClone. You can use built-in characters, your own assets, or the default options in AI Studio.

• Pose actors: These are static characters with pre-made poses. Drag them into the scene to quickly block out compositions and storytelling moments.
• Motion tools: Use Edit Motion Layer to fine-tune body movement and Face Key to adjust facial expressions for more nuanced performances.
• Props: A large library of buildings, objects, vehicles, and more lets you rapidly build environments with different moods and styles.

Because the 3D setup is the foundation, you can iterate quickly: change a pose, tweak a facial expression, or move a prop, then regenerate with AI while keeping overall consistency.

Generating AI images from the 3D viewport

Once your character and scene are in place, you can capture the current camera view and send it to AI Studio’s 3D-to-image workflow.

• Position your character and apply a pose.
• Click Capture Image from Viewport to send that exact framing to AI Studio.
• Choose an AI model and style, then generate.

Choosing AI models and aspect ratios

Different AI models produce different looks:

• Flux 2 Max: Smooth skin, cinematic, slightly younger-looking characters.
• Nano Banana Pro: More detailed, realistic, often older-looking characters.

You can also set the output aspect ratio (for example, 3:4 or 16:9). When you change the ratio in AI Studio, the iClone viewport automatically updates to match, so your composition stays accurate.

Using prompts, lighting, and presets

AI Studio includes prompt presets for quick experimentation:

• Inspiration presets: Combine scene, mood, and lighting ideas in one go.
• Lighting presets: Focus on light direction, softness, and atmosphere.
• Style presets: Control the overall visual look (photorealistic, stylized, etc.).

For characters, context matters. A simple prompt like “white background, soft light” can turn a plain 3D viewport into a convincing studio portrait, with unified colors and realistic lighting.

Keeping faces consistent with AI actors

To maintain a stable character identity across multiple images, AI Studio offers AI actors—pre-trained character profiles designed for facial consistency.

• Click Add Actor to browse over 200 AI actors.
• Pick one that’s visually close to your 3D character (for example, Camilla or Kevin).
• The actor’s tag is automatically added to your prompt to lock in that identity.

If you accidentally remove the tag, you can restore it with Add Tag. With AI actors, you can change camera angles, poses, and lighting while still getting a face that looks like the same person every time.

Controlling backgrounds with HDRI environments

Prompt-only backgrounds can be unpredictable. To get more control, you can use HDRI environments as your base.

• In AI Studio, select a Stage and choose from many built-in HDRIs (indoor, outdoor, studio, etc.).
• Apply the HDRI, then click Capture Image from Viewport and add a short descriptive prompt.

The AI will respect the HDRI as the environmental reference, giving you strong consistency between character and background across multiple generations.

Fine-tuning HDRI lighting and orientation

Under Image-Based Lighting (IBL) settings, you can rotate the HDRI to change the visible background and light direction.

• Use Rotate IBL to spin the environment until the composition looks right.
• Capture and regenerate to see the updated angle while maintaining continuity.

You can also load your own HDRIs, such as an exhibition hall or showroom. From there, drop in props (like cars) and pose your character to match the story—e.g., Camilla inspecting a car in a dealership. The AI blends character, prop, and HDRI into a coherent scene.

Using dummy objects to design layouts with prompts

Dummy objects are simple colored shapes you place in 3D space to represent future furniture or props. You then tell the AI what each color means in your prompt.

For example:

• Red dummy = sofa
• Green dummy = wooden table
• Blue dummy = pouf

By describing this mapping in your prompt, you can control where each object appears and what material or style it should have. This works for both characters and props, letting you build complex layouts without modeling every detail.

Changing outfits and refining performances

You’re not limited to a single look for your character. You can:

• Use prompts to request different outfits that match the scene (e.g., casual clothes for car shopping).
• Or swap clothing directly in the iClone content manager before generating.

Combine this with motion editing and facial keys to refine posture, gestures, and expressions. When you regenerate, the AI preserves motion and expression while updating style, clothing, or background.

Turning 3D scenes into AI-generated videos

Once you’re happy with a still image, you can move into 3D-to-video generation.

• Go to the Video tab and click Create—your last image is loaded as the starting frame.
• Choose a video model such as Seedance 2. It costs more credits but understands prompts and motion descriptions more accurately.
• Describe the character’s actions, context, and any dialogue in the prompt.

If you include spoken lines, enable Generate with Audio so the video is created with matching narration.

Adding camera presets and narration

AI Studio offers camera presets like dolly in, where the camera slowly moves closer to the character. You can set resolution and duration, then generate a clip where the character moves and speaks according to your description.

This is ideal for cinematic character moments—like a hero walking toward the camera while delivering a key line.

Animating complex motions with 3D references

Some actions are too complex to describe well with text alone, especially interactions or crowds. That’s where iClone’s motion tools shine.

• Use Edit Motion Layer to create or refine complex movements (fights, interactions, gestures).
• Capture the animation with Capture Video from Viewport so AI Studio gets a full motion reference, not just a static frame.

In the prompt, you can still add extra details that don’t exist in 3D, such as “a large crowd in the background, crowds waving hands”. The AI respects the 3D motion for your main character while generating background motion based on your description.

Designing custom camera movement

Camera work is a big part of cinematic storytelling, and 3D guidance gives you full control.

• Manually animate the camera in iClone during a sequence (for example, a fight scene).
• Ensure characters stay in frame while you add dramatic moves like sweeps, pushes, or tilts.
• Capture the video from the viewport and enable Apply Camera Movement in AI Studio.

When you generate with Seedance 2, the AI keeps both character motion and camera animation highly consistent, even if you also add environmental prompts like “heavy rain” or specific visual styles.

Using real camera presets

In the content manager under Stage > Camera, you’ll find real camera presets that mimic common cinematography techniques:

• Spiral
• Pan
• Orbit
• Truck and more

Apply a preset, preview the move, then send it to AI Studio. The AI respects the exact path and timing of the camera, so you can plan shots like a real director.

Creating multi-shot sequences with camera switching

For more advanced storytelling, you can build multi-shot edits directly in iClone using multiple cameras.

• Set up Camera 1 for a wide establishing shot.
• Create Camera 2 for a mid shot.
• Add Camera 3 for a close-up on the character’s face.

Then use Switch Camera and add a switcher track:

• Assign Camera 1 at the first keyframe.
• Add Camera 2 at the next cut point.
• Add Camera 3 at the final cut point.

On playback, the system automatically cuts between cameras along the timeline. After capturing the animation, you can combine it with a first-frame image—for example, a 19th-century restaurant—and describe the scene and dialogue in your prompt.

The result is a fully AI-rendered multi-shot sequence with consistent characters, motion, and environment, driven by your 3D direction.

Why 3D guidance is a game changer for AI visuals

By using iClone’s 3D tools as the “director” and AI Studio as the “visual engine,” you can achieve results that are almost impossible with text prompts alone:

• Reliable character appearance and facial identity across shots
• Accurate, repeatable poses and animations
• Controlled environments using HDRIs and props
• Professional camera work, including complex moves and multi-shot edits

If you’re interested in pushing AI visuals even further, you might also like exploring how to direct a cinematic AI music video with another tool like Hailuo MiniMax, or how modern tools can turn simple images into production-ready 3D models that you can bring back into scenes like these.

With this 3D-guided workflow, you’re no longer at the mercy of random generations. You can plan, block, and direct your scenes like a filmmaker—then let AI handle the heavy lifting of rendering style and detail.