StudyUlt AI Study OS

A localhost-first, markdown-native AI study operating system for JEE, NCERT, boards, active recall, and graph-based learning.

Run Locally

npm install
npm run dev

Open http://localhost:3000.

What Is Built

• Next.js, React, TypeScript, TailwindCSS
• Framer Motion animated dashboard and transitions
• Three.js particle backdrop through React Three Fiber
• Cytoscape.js Obsidian-style concept graph
• Markdown vault parser with frontmatter, tags, wikilinks, backlinks, formulas, flashcards, and tests
• MiniSearch instant fuzzy search
• LaTeX rendering with remark/rehype
• Local AI route with pluggable provider fallback for Ollama, LM Studio, OpenRouter, Claude, Gemini, and DeepSeek style endpoints
• Project MCP setup for 21st.dev Magic, Magic UI, Shadcn, Aceternity registry, and React Bits registry

Vault Structure

Markdown lives in MasterVault.

MasterVault/
  Physics/
  Chemistry/
  Mathematics/
  Flashcards/
  Tests/
  Analytics/

Any .md file is indexed on API refresh. Use:

---
title: Gauss Law
tags: [JEE, Physics, weak]
---

Link to [[Electric Field]] and formulas like $E = kq/r^2$.

- Q:: Question text A:: Answer text
- MCQ:: Prompt Options:: A | B | C | D Answer:: A

MCP And Registries

Project-local MCP config is in .mcp.json and .vscode/mcp.json.

Shadcn registry aliases are in components.json:

• @aceternity
• @react-bits

Codex MCP server entries were also added to C:/Users/Rehan/.codex/config.toml for this trusted workspace.

AI Providers

For local inference, set LOCAL_AI_ENDPOINT in .env.local. The API route posts:

{
  "provider": "local",
  "prompt": "Explain Gauss Law",
  "context": []
}

Without an endpoint, the app uses a deterministic local tutor fallback connected to the vault analytics.