Open Canvas →

Canvas Guide

Everything you can do on the canvas right now. Start at the top, work your way down. Each workflow has copy-pasteable content so you can test immediately.

How the canvas works
Blocks — everything on the canvas is a block. Text, URLs, images, chat, code, notes. Click a type in the Library sidebar to create one.
Interactive — every block has inputs. Click into a text block and type. Paste a URL and press Enter. Drop an image file.
Actions — most blocks have action buttons at the bottom. "Summarize" creates a new card with an AI summary. "Generate image" creates an image from your text. "Read aloud" plays TTS.
Proximity = context — Chat blocks read nearby content automatically. Place things near each other to create context. Move them apart to separate contexts.
Persists — your canvas saves automatically. Close the tab, come back, everything is there.
💬Start Here~30 sec
Your First Canvas Chat
Place a Chat block on the canvas and talk to AI.
  1. 1Open the canvas (button above)
  2. 2In the Library sidebar on the left, click "Chat" under AI
  3. 3A chat card appears — click into the text input at the bottom
  4. 4Type "Hey, what can I do on this canvas?" and press Enter
  5. 5Watch the AI stream a response inside the card
  6. 6Try the suggestion chips in an empty chat to get started fast
Why this matters
Chat lives on the canvas as a spatial object. You can place multiple chats, move them around, and they pick up nearby content as context.
📝Start Here~30 sec
Write on the Canvas
Create a Text block and start typing — it saves automatically.
  1. 1Click "Text" in the Library sidebar
  2. 2A text card appears — click into it and start typing
  3. 3Everything you type is saved automatically
  4. 4Try the action buttons at the bottom: "Summarize", "Generate image", "Read aloud"
Why this matters
Every block is interactive. Click and type — no edit mode, no double-click ritual. Content persists across reloads.
🔗Start Here~1 min
Scrape a Web Page
Add a URL block, paste a link, and watch it fetch the content.
  1. 1Click "Web page" in the Library sidebar
  2. 2A card appears with a URL input — paste one of the URLs below
  3. 3Press Enter or click "Go"
  4. 4Watch the skeleton loading animation as it scrapes
  5. 5Content fills in: title, excerpt, full text
  6. 6Click "Summarize" to get an AI summary in a new card
Copy & paste
Why this matters
Any URL becomes a live card with extracted content. The AI can read it, summarize it, generate images from it.
▶️Start Here~1 min
Get a YouTube Transcript
Paste a YouTube link and get the full transcript + summary.
  1. 1Click "YouTube" in the Library sidebar
  2. 2Paste a YouTube URL below into the input
  3. 3Press Enter — watch the loading skeleton
  4. 4The transcript and summary fill in
  5. 5Click "Summarize" for key points, or "Mindmap" for a structured outline
Copy & paste
Why this matters
Video content becomes text you can work with. Summarize, mindmap, ask questions about it.
🧲Explore~2 min
Proximity = Context
The core mechanic: place content near a Chat block and the AI uses it.
  1. 1Create a Text block and type: "Our startup has 10K users, $50K MRR, and we sell AI design tools."
  2. 2Create another Text block nearby: "Our competitor just raised $20M and has 50K users."
  3. 3Place a Chat block between them
  4. 4Ask: "What should our strategy be?"
  5. 5The AI references both text blocks — it sees what's near it
  6. 6Now drag one text block far away (zoom out, move it to a distant area)
  7. 7Ask the same question again — the AI no longer references the distant card
Why this matters
You compose AI context by arranging things spatially. No wiring, no configuration. Near = related. Far = separate.
💻Explore~1 min
Write & Store Code
Create a Code block with a language selector and monospace editor.
  1. 1Click "Code" in the Library sidebar
  2. 2A code card appears with a monospace editor
  3. 3Click the language dropdown (top right) to select JavaScript, Python, etc.
  4. 4Type or paste code — it's always editable
  5. 5Place a Chat block nearby and ask "Explain this code" — the AI reads it via proximity
Why this matters
Code lives on the canvas alongside your notes, research, and conversations. The AI can read and discuss any code block near it.
📌Explore~30 sec
Quick Notes
Drop sticky notes anywhere for quick thoughts.
  1. 1Click "Note" in the Library sidebar
  2. 2A warm yellow card appears — just start typing
  3. 3Use notes to annotate your canvas, mark ideas, or leave reminders
  4. 4Notes are simple on purpose — no action buttons, just text
Why this matters
Sometimes you just need a sticky note. Not everything needs AI.
🖼️Explore~30 sec
Drop an Image
Drag an image file onto the canvas or paste an image URL.
  1. 1Click "Image" in the Library sidebar
  2. 2A drop zone card appears
  3. 3Drag any image file from your computer onto the card
  4. 4Or type an image URL into the input and press Enter
  5. 5The image renders inside the card
Why this matters
Images are first-class content on the canvas. Place them near a Chat and ask the AI about them.
🔬Power User~3 min
Multi-Source Research
Paste multiple articles, then synthesize with a Chat block that sees all of them.
  1. 1Create 3 Web page blocks and paste the URLs below into each
  2. 2Wait for all 3 to finish loading (skeleton → content)
  3. 3Place a Chat block in the center of all 3 cards
  4. 4Ask: "What are the key themes across these articles? Where do they agree or disagree?"
  5. 5The Chat reads all 3 via spatial proximity and synthesizes
  6. 6Click "Summarize" on the Chat response for a tighter version
Copy & paste
Why this matters
Multi-source synthesis that would take 30 minutes of tab-switching happens in 3 minutes on one canvas.
🎨Power User~2 min
Text to Image
Write a description, click "Generate image", and get an AI image.
  1. 1Create a Text block and write: "a cozy library with sunlight streaming through stained glass windows"
  2. 2Click "Generate image" in the action bar at the bottom
  3. 3A new Image block appears to the right — wait for it to generate
  4. 4The prompt is auto-enhanced before generation for better results
  5. 5Try another: "underwater city with bioluminescent buildings"
  6. 6Compare both images on your canvas
Why this matters
Text becomes images. The output is a canvas card you can move, compare, and build on.
⛓️Power User~4 min
Content Chain
YouTube → summary → blog post → image. One input, many outputs.
  1. 1Create a YouTube block with the URL below
  2. 2Wait for the transcript to load
  3. 3Click "Summarize" → summary card appears
  4. 4Place a Chat block near the summary
  5. 5Ask: "Write a 200-word blog post based on this summary"
  6. 6Click "Generate image" on the blog post for a header image
  7. 7Click "Read aloud" on any text to hear it via TTS
  8. 8You now have: transcript → summary → blog → image → audio
Copy & paste
Why this matters
Each block's output feeds the next. The canvas is a content multiplication machine.
⚖️Push It~4 min
AI Debate
Three Chat blocks debate a topic using spatial context.
  1. 1Create a Text block: "Should AI models be open-sourced?"
  2. 2Place Chat A to the left of the text. Ask: "Argue strongly FOR open-source AI with specific examples."
  3. 3Place Chat B to the right. Ask: "Argue strongly AGAINST open-source AI with specific examples."
  4. 4Wait for both responses
  5. 5Place Chat C below all three, within proximity of both A and B
  6. 6Ask C: "Based on the arguments near you, who makes the stronger case? What's the synthesis?"
  7. 7Chat C reads both sides via proximity and judges
Why this matters
Multiple AI perspectives, spatially arranged, cross-referencing each other. Collective intelligence on a canvas.
🚀Push It~5 min
Speed Learn a Topic
How fast can you learn something new using the canvas?
  1. 1Pick a topic you know nothing about
  2. 2Create 3 Web page blocks with relevant Wikipedia links (examples below)
  3. 3While they scrape, create a YouTube block with a relevant video
  4. 4Click "Summarize" on each completed card
  5. 5Place a Chat block in the center of everything
  6. 6Ask: "Give me a beginner-friendly explanation of [topic] in 3 paragraphs"
  7. 7Ask follow-up: "What are the 3 most important concepts to understand?"
  8. 8Click "Mindmap" on any summary for a structured outline
Copy & paste
Why this matters
If the canvas makes learning 10x faster, it's a daily-use product. This is the acid test.
🌱Push It~10+ min (ongoing)
Knowledge Garden
Build a persistent knowledge space. Come back tomorrow and keep growing it.
  1. 1Pick something you're genuinely working on or curious about
  2. 2Create a Text block as the root: your central question or goal
  3. 3Add content: paste URLs, YouTube videos, your own notes
  4. 4Use Chat blocks to ask questions and get summaries
  5. 5Arrange everything spatially — group related items together
  6. 6Use Note blocks to mark areas: "Readings", "Ideas", "Questions"
  7. 7Zoom out — you should see the shape of your thinking
  8. 8Close the tab. Come back tomorrow. Everything is still there.
  9. 9Add new things. The garden grows.
Why this matters
This is the real test. If it feels natural to come back and build on your canvas over days, the product works.

Block Types Reference

💬Chat
AI conversation with spatial context
📝Text
Editable text, always live
🔗Web page
Paste URL → auto-scrapes content
▶️YouTube
Paste link → transcript + summary
🖼️Image
Drop file or paste URL
💻Code
Monospace editor with language select
📌Note
Simple sticky note

Action Buttons Reference

📋
Summarize — Creates a new text card with an AI summary of the block content
🎨
Generate image — Creates a new image card from the text content (auto-enhances the prompt)
🧠
Mindmap — Creates a structured outline/mindmap from the content
🔊
Read aloud — Plays the text as speech via TTS
While testing, notice...

Does clicking into blocks feel instant? Can you type without friction?

Do the loading skeletons give you confidence something is happening?

When a new card appears from an action, does the position feel right?

Does spatial context actually work? Does the Chat reference nearby content?

After 10 minutes, does your canvas feel organized or chaotic?

Would you come back tomorrow to keep building on this canvas?