Skip to content

om-design/mai-canvas

mai-canvas

A personal modular dashboard linking personal content and AI services


Vision

mai-canvas is an open-source platform that puts you in control of your digital life.
Visually connect your files, notes, and personal assets to local and external AI services on a modular canvas—always knowing what stays private and what leaves your device.

  • Yellow Border: Instantly see which data and services are local (inside the border) and which are external (outside).
  • Wiring: Drag and drop connections between your content and AI tools, with clear visual cues for privacy and data flow.
  • Privacy by Design: Data only leaves your device when you explicitly allow it. Use vector abstraction to leverage powerful AI without exposing your raw content.
  • Composable: Mix and match local and external models, automate workflows, and chain together multiple agents for richer, more reliable results.
  • Extensible: Add new services, plugins, and connection types as the ecosystem grows.

Why?

  • User Control: You decide what happens with your data, every step of the way.
  • Privacy: Keep sensitive information local, use abstraction for safe external processing.
  • Modularity: Build your own workflows, combine the best tools, and adapt as AI evolves.
  • Transparency: Visualize data flow and service connections at a glance.

How It Works

  • Canvas UI: Drag-and-drop interface with a yellow border marking your private zone.
  • Blocks: Represent local assets, AI models, and external services as modular blocks.
  • Wires: Connect blocks to create workflows. Wires crossing the yellow border indicate data leaving your device.
  • Abstraction Services: Use vector embeddings to perform useful work (search, clustering, recommendations) without sharing raw content.
  • Multi-Agent Chaining: Run tasks through multiple models or services for consensus, verification, or multi-perspective results.

Get Involved

We’re looking for:

  • Developers (frontend, backend, AI, privacy)
  • Designers (UI/UX, workflow visualization)
  • Privacy and security experts
  • Early adopters and testers

Join us to build the future of personal AI!


License

MIT License


Roadmap

  • Canvas UI prototype (web-based)
  • Local asset management and yellow border logic
  • Pluggable service and connection library
  • Vector abstraction service (privacy-preserving)
  • Multi-agent chaining and consensus
  • Documentation and community onboarding

Inspiration

Node-RED, n8n, Pinokio, Agregiore, Smithery, Obsidian, Logseq, OpenMined, and the emerging MCP ecosystem.


Contributing

See CONTRIBUTING.md for guidelines, or open an issue to start a discussion!


Let’s build a personal AI OS that puts you in control—private, modular, and future-proof.

About

A personal modular dashboard linking personal content and AI services

Resources

License

Code of conduct

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages