Generate: "create 5 cat images" Edit: Paste an image onto the canvas, then say "make @i3 pink" Reference: Use @i/@v/@t to reference images, videos, and notes (e.g., "make @i3 pink" or "animate @i5 into a video") Notes: Click "📝 Create Text Node" button or say "create a note" Canvas: Images automatically appear on canvas with IDs!
Welcome to Canvas Agent
Get started in 3 steps:
1Add your Gemini API key
2Paste or drop an image onto the canvas
3Type what you want: "make @i1 more colorful"
🚀 Try a Workflow
Click to load example images and see how it works:
Full Capabilities
🎨 Image Generation & Editing
generate 5 cat imagesCreate new images from text
make @i3 pinkEdit existing images by ID
generate in 4KHigh resolution output (2K, 4K)
create a 16:9 landscapeCustom aspect ratios
🔗 Reference Modifiers
@i3:styleUse image as style reference only
@i5:base @i7:styleEdit base image with another's style
@i1:base add @i2 to itComposite images together
🎬 Video Generation (Veo 3)
create a sunset videoText-to-video generation
animate @i1Turn any image into video
@i1 first frame @i2 last frameTransition between two images
describe @v1Analyze video content
🔊 Audio Generation (Fish Audio)
generate audio "Hello world"Text-to-speech
create narration for scene 1Voice narration
📖 Story Mode
create a 5 scene story about...Auto-generate full visual stories
story for youtube shortFormat-specific stories
📝 Notes & Organization
create a note with tasks...Add text notes to canvas
@t1Reference notes in prompts
🧩 Templates & Workflows
/template_nameUse saved prompt templates
Create custom templates for reusable prompts
Load workflow presets with example images
⚡ Smart Autocomplete
@Quick reference images, videos, audio, notes
/Browse and insert templates
:Add modifiers (:style, :base)
🖼️ Canvas Features
Paste images directly (Ctrl+V)
Box-select multiple items
Pan mode (Space or H)
Export/Import canvas as JSON
Multiple sessions with unique URLs
Create Template
Saved Templates
Template Preview
Session Information
Session Tips:
Each session has its own canvas and conversation history
Sessions are automatically saved and isolated from each other
Share sessions by copying the URL or session ID
Create multiple sessions to organize different projects
Workflows
Select a workflow template to load example images and a sample prompt onto your canvas.
Flash 2.5 is the default for speed; Pro 2.5 adds extra reasoning depth; Gemini 3 Pro offers the most advanced reasoning.
Gemini 2.5 Flash Image is fast and efficient; Gemini 3 Pro Image Preview offers higher quality, 4K resolution, and advanced features like Google Search grounding.
Free Trial Usage
Agent
0/20
Image
0/5
Free trial: 20 agent + 5 image requests per day. Add your API key for unlimited use.