The AI Image Generator lets you create images for your storyboard frames using text prompts. Reference images are the most powerful way to maintain consistency across your storyboard, allowing you to carry characters, locations, and visual elements from frame to frame.
This guide covers how to use the generator effectively, from basic prompts to advanced techniques.
Before generating images, select a visual style that matches your project. Click the Style dropdown to see available options.

Each style produces distinctly different results—preview the style thumbnails to find one that suits your needs. You can change styles at any time and regenerate images.
Tip: Experiment with different styles early in your project to find the best match for your creative direction.
Reference images help maintain consistency across your storyboard. You can use existing frame images, generate images in the playground, or upload your own assets as references.
Important: The AI doesn't automatically remember previous frames. Writing "the same character" or "the location from frame 3" in your prompt won't work—you must attach the reference image for the AI to see it.

Upload and use your own images as references, or use assets generated in the playground:
| Feature | Description |
|---|---|
| Upload images | Add character designs, location references, or style guides |
| Reuse across frames | Same asset available for all frames in your storyboard |

Use existing storyboard frames as references:
| Feature | Description |
|---|---|
| Maintain consistency | Reference a frame to keep characters or settings consistent |
| Build sequences | Use the previous frame to maintain visual continuity |
| Best for | Continuing scenes, character consistency, location matching |

You can attach multiple reference images to combine elements:
When using reference images, you can choose between two generation modes:

| Mode | Speed | Quality | Availability |
|---|---|---|---|
| Fast | 15-30 seconds | Good for rapid iteration and exploration | All users |
| Precise | 45+ seconds | Higher fidelity, better prompt following and style matching | Paid plans only |
Use Fast mode when exploring ideas and iterating quickly. Switch to Precise mode when you need the highest quality output and closer adherence to your references.
The generator input field changes color to indicate its current state and what type of generation will occur.

A blue outline indicates standard generation mode. The AI will create an image using your prompt text and selected style settings. This is the default state when no reference images are attached.

A yellow background indicates that reference images are attached. The AI will use these references along with your prompt to generate an image that incorporates elements from your reference images.

A red background indicates a problem: you've attached a reference to another frame, but that frame doesn't have an image yet. You need to generate or upload an image to the referenced frame first, then return to continue.
When your frame has text content (like action or dialogue fields), press Tab in an empty prompt field to automatically generate a prompt based on that content.

This feature:
Tip: Tab autocomplete is a fast way to generate images that match your script content.
The prompt describes what you want to see in your image. Good prompts are specific and descriptive.
| Approach | Example |
|---|---|
| Be specific | "A detective in a trench coat examining a clue under a streetlight" rather than "a person" |
| Include details | Describe lighting, mood, colors, and environment |
| Mention important elements | Characters, objects, settings, and actions |
Note: Avoid phrases like "the same as before" or "from frame 5" in your prompts. The AI only sees what you attach as reference images—it doesn't have context about your other frames unless you explicitly attach them.

When working with reference images, start your prompt with "Show me" for better results:
| Prompt type | Example |
|---|---|
| Without prefix | "The character walking through a forest" |
| With prefix | "Show me the character walking through a forest" |
The "Show me" prefix helps the AI understand you want to see the referenced content in a new context.
Important: When using reference images, remember to start your prompt with "Show me" for best results.
When generating images without reference images attached, you can select a shot type to control the camera framing. Click the Shot type dropdown to choose from available options.

| Shot type | Description |
|---|---|
| Auto | AI determines the best framing |
| Close up | Tight framing on subject's face or details |
| Mid shot | Subject framed from waist up |
| Wide shot | Full subject with surrounding environment |
| From below | Camera looking up at subject |
| From above | Camera looking down at subject |
| Aerial | Bird's eye view from directly above |
Note: The shot type selector is only available when no reference images are attached. When using references, specify your desired camera angle in the prompt instead (e.g., "Show me a close up of the character").
The AI Image Generator uses a credit system:
| Action | Credit cost |
|---|---|
| Generating preview variations | Free |
| Inserting an image into your storyboard | 1 credit |

You can generate as many preview variations as you like without using credits. Credits are only consumed when you choose to insert an image into your frame.
Note: Workflow plan subscribers have unlimited image credits.