Tools
Every agent tool, what it does, and what you need selected to use it.
Tools live in the agent input on the board. Some require a specific selection on the canvas. Others, like the libraries, open the unified Libraries dialog launched from the chat input.
When you select nodes on the board, only the tools that match your selection appear as quick actions.
Agent tools
| Tool | What it does | What you select |
|---|---|---|
| Ask Agent | Ask or instruct the agent about selected nodes. | One or more nodes. |
| Describe | Analyze and describe any visual content. | One or more image or video nodes. |
Image tools
Image generation uses two mode-aware quick tools plus the existing utility tools. Edit Image modifies an existing image; Image to Image uses selected images as style or composition references for a new image. Each tool filters the model picker to only models that support the chosen mode.
| Tool | What it does | What you select |
|---|---|---|
| Edit Image | Modify or incorporate elements of an existing image. Filters the model picker to edit-capable models. | A frame or image. |
| Image to Image | Generate a new image guided by one or more reference images. Filters the model picker to reference-capable models (includes Reve and Ideogram V3). | One or more images as references. |
| Combine Images | Generate a new image by combining elements of two source images. | Exactly two image nodes. |
| Image Upscale | Enhance image resolution and quality up to 4K. | One or more images. |
| Change Aspect Ratio | Regenerate at a different aspect ratio. | A frame or image. |
| Remove Background | One-click background removal. | One or more images. |
| Enhance Texture | Add detail and sharpness to an image's surfaces and textures. | One or more images. |
Text-to-image generation doesn't appear as a quick action — start it by describing the image in the agent input without a selection.
Video tools
Video generation is split into three mode-aware tools plus two "replace" shortcuts. The tool you see in the quick-action menu depends on what you have selected.
| Tool | What it does | What you select |
|---|---|---|
| Image to Video | Animate a single image into a short video clip. | Exactly one image (used as the start frame). |
| Frames to Video | Generate a video that transitions between two keyframes. | Exactly two images (start and end frames). |
| Reference to Video | Generate a video guided by images and/or videos used as style references. | One or more images or videos as references. |
| Replace in Image | Replace a subject or element inside an image using a second image as the reference. | Exactly two images. |
| Replace in Video | Replace a subject or element inside a video using an image as the reference. | Exactly one image plus one video. |
| Video Upscale | Enhance video resolution and quality up to 4K. | A video node. |
| Apply Motion | Transfer movements from a reference video to a character image. | A character image plus a movement reference video. |
| Multi-Scene Video | Combine multiple images into a storyboard-style video. | Four or more keyframe images. |
| Join Clips | Join two video clips with a transition. | Exactly two video clips. |
Text-to-video generation doesn't appear as a quick action — start it from the Clapperboard entry in the toolbar or by describing the video in the agent input without a selection.
Animation
| Tool | What it does | What you select |
|---|---|---|
| Animate | Animate layers within a frame. | One or more frames containing at least 2 layers. |
| Edit Video Tracks | Ask the agent to edit existing video tracks — animation duration and existing tracks are read and updated coherently with video tracks. | A frame with video tracks. |
Inspection
| Tool | What it does | What you select |
|---|---|---|
| Inspect | View node type, dimensions, aspect ratio, and more. | One or more nodes. |
| Inspect Video | Returns a camera-motion prompt, a video-level caption, and dense per-clip event captions so the agent can reason about the footage. | One or more video nodes. |
| Inspect PDF | Analyze a PDF asset — ask questions about it in chat or have the agent generate content based on its contents. | A PDF asset. |
Libraries
Libraries open the unified Libraries dialog from the chat input rather than prefilling the agent.
| Library | What it contains |
|---|---|
| Brand Assets | Logos, fonts, images, and brand materials. |
| Mockups | Template mockups for inspiration. |
| Camera Motion | Curated camera-motion presets for video. |
| Easing | Easing curves you can pair with a camera motion. |
| Video Transitions | Transition effects for joining video clips. |
A new image-style library is in development to replace the sunsetted LoRA training flow.