About this project

Nesso allows users to build knowledge graphs where nodes represent ideas and edges represent typed semantic relations (featuring 52 types across 8 categories). To support active learning, it integrates the Free Spaced Repetition Scheduler (FSRS) to manage the review state of concepts. Key features include: - Multi-platform availability: Accessible as a web application or as native desktop builds for macOS and Windows via Tauri. - AI Integration: An optional AI mentor named Socrates that connects to OpenAI-compatible endpoints (including local Ollama) to interact with the live graph. - Data Persistence: Uses IndexedDB for web storage and plain JSON files for desktop project storage. - Modular Architecture: Built as a pnpm monorepo with separate packages for schema management, vocabulary learning, design tokens, an embeddable React graph component, and an MCP server for LLM clients. The application is developed using React 18, Vite, TypeScript, and React Flow for the canvas rendering.