Create consistent AI-generated images
The Generate sidebar is the primary surface for creating and editing storyboard frame images in Boords. Open it from any frame to sketch a layout, prompt a new frame, edit part of an existing image, change the artwork style, or reframe the shot. Every generation is saved as a frame revision, so you can page back through the history at any time. See Frame revisions for how revisions work.
Reference images are the most powerful way to maintain consistency across your storyboard, letting you carry characters, locations, and visual elements from frame to frame. This guide covers how to use the Generate sidebar effectively, from basic prompts to advanced techniques.
Open any frame and switch to the Generate tab in the sidebar. Generate has three sub-tabs:

This guide focuses on Canvas. See Styles & restyling for Restyle, and the Camera sub-tab section below for reframing.
Canvas is the main generation surface. It works in three modes, switched by the Source picker at the top of the panel.
| Source | What it does |
|---|---|
| Blank | Generate a new image from scratch using your brush layout and prompt |
| Current | Edit the current image in this frame by brushing over the region to change, then prompting the change |
| Another frame | Use another frame's image as a reference point for generation |

The Canvas toolbar gives you a brush, size and colour controls, and a prompt field. Block out the composition with rough shapes (stick figure for the actor, rectangle for the window, circle for the cup of coffee) and write a prompt describing what you want. The brush gives the AI a layout to follow.
Add a prompt like "woman in her thirties working at a sunlit desk, soft morning light from the left" and click Generate. The result keeps your composition while filling in the prompt's details. Iterate the prompt without losing the layout.

Set the Source to Current to edit the image already in the frame. Brush over the part you want to change, then prompt the change. Boords focuses the work on the brushed region and keeps everything outside it close to the original. The brush mark is a hint, not destructive. It's discarded after the request.
Useful examples:
The previous version is saved as a frame revision, so you can page back if the change doesn't land.

The Camera sub-tab takes the current image and regenerates it from a different angle, height, or distance. Pick a preset and Boords renders the same shot from the new viewpoint. Use it when the composition is right but the camera position isn't.

Each reframe is saved as a frame revision, so you can compare angles side by side and revert if a take doesn't land.
Reference images help maintain consistency across your storyboard. You can use existing frame images, generated assets, or your own uploaded files 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:
Pick 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. Some styles also support a custom color palette for brand consistency.
Tip: Experiment with different styles early in your project to find the best match for your creative direction. See Styles & restyling for the full list and colour palette controls.
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 an image in Canvas 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").
For reframing an existing image to a new angle, see the Camera reframe section above.
The Restyle sub-tab changes the artwork style of the current image without rebuilding the scene. The layout, characters, and poses stay intact while the style swaps over.
See Styles & restyling for the full walkthrough.
Every Canvas generation, Camera reframe, Restyle, drag-dropped local image, and inline grid generation creates a frame revision. Revisions survive reload, ride with the frame, and can be paged through, locked, or deleted. See Frame revisions for the full guide.
The 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: Your monthly credit allowance depends on your plan. Pro gets 1,000/mo, Team gets 2,000/mo, and Agency gets 3,000/mo. See Billing & plans for details.
The storyboard grid also has a quick generator for first-pass populating frames. Use it when you want to fill a fresh storyboard fast, then refine individual frames from the Generate sidebar. Each grid generation saves as a frame revision, so any sketch, restyle, or edit you make later keeps the history.
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. Boords 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. Boords 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.