A canvas where your AI builds the dashboard.
awareness-canvas is the visual companion to mcp-awareness. Instead of a fixed dashboard with pre-built widgets, you have a chat panel and a spatial canvas. Ask your AI for a visualization and it generates a React component on the fly, queries Awareness for the data, and renders it as a draggable, resizable widget on your canvas.
Over time, your canvas becomes your personalized dashboard — not because you configured it, but because you asked for what you wanted and kept what was useful.
Planning phase. See docs/design.md for the architecture and approach.
- You type in the chat panel: "Show me my pending intentions on a timeline"
- The AI generates a self-contained React component that queries Awareness
- The widget appears on the canvas — drag it, resize it, keep it or close it
- Say "add a tag cloud next to it" and another widget appears
- Your layout persists across sessions
- No widget catalog to learn. You describe what you want in natural language.
- Widgets are data-aware from birth. The AI generates them with the query baked in.
- The AI knows your data shape (via
get_stats,get_tags) and can suggest visualizations. - Widgets can be conversational. "This chart looks wrong, filter out the test entries" and the AI regenerates it in place.
- Some core widgets ship pre-built for zero-friction onboarding.
mcp-awareness is the store — knowledge, intentions, alerts, patterns. It speaks MCP.
awareness-canvas is the surface — a web app that talks to Awareness through the AI. The AI is the API layer. The canvas never touches the database directly.
Part of the Awareness ecosystem. Copyright (c) 2026 Chris Means.