A spatial research tool designed to augment human thinking with AI, moving beyond traditional chat interfaces. It targets researchers and power users by enabling AI to quietly classify, connect, and synthesize notes on a canvas, enhancing associative thinking without dominating the user's workflow. The benefit lies in a more integrated and less intrusive AI assistance model.

How It Works

Notes are placed on a dynamic spatial canvas where AI automatically classifies them into 14 types, enriches them with annotations, and infers content-based connections. This approach prioritizes spatial and associative thought, with AI operating subtly in the background. When sufficient notes accumulate, the system synthesizes emergent insights, which can be solidified by the user. Users can switch between Tiling (BSP grid), Kanban (type-grouped), and Graph (force-directed) views.

Quick Start & Requirements

• Install/Run: Clone the repository, run npm install, then npm run dev. Access via localhost:3000.
• Prerequisites: A desktop browser and an OpenRouter API key are mandatory. The key is stored client-side in localStorage.
• Models: Supports various models including GPT-4o (default), Claude-Sonnet-4.5, Gemini-2.5-pro (with web grounding), DeepSeek-Chat, and Mistral-Small-3.2.
• Optional: Web grounding requires models supporting the :online suffix.
• Resources: An introductory video is linked within the README.

Highlighted Details

• AI classifies notes into 14 distinct types (e.g., claim, question, idea, entity).
• Connection indicators visually highlight relationships between notes.
• Emergent synthesis feature generates single-sentence insights from canvas content.
• Data is entirely client-side, stored in browser localStorage with silent rolling backups.
• Supports export to .md or .nodepad (JSON) formats.
• Built with Next.js, React 19, TypeScript, Tailwind CSS v4, D3.js, and Framer Motion.

Maintenance & Community

This project appears to be a personal design experiment by Saleh Kayyali, with no explicit mentions of a broader community, roadmap, or formal maintenance structure in the provided README.

Licensing & Compatibility

The README does not specify a software license, making its terms for use, modification, and distribution unclear. This lack of licensing information poses a significant adoption blocker for commercial or collaborative projects.

Limitations & Caveats

Data persistence is limited to browser localStorage, which may not be suitable for very large projects or collaborative workflows. As a "design experiment," the project may be subject to rapid changes, potential instability, or incomplete feature sets. The absence of a clear license is a critical caveat.