NoteCapt — turn a messy pile of study materials into a clean, AI-readable Markdown library (macOS)

[kasingbbb]

NoteCapt — a macOS study-materials workbench, and my first publicly shipped Tauri app.

Drag in PDFs, slides, EPUBs, images, audio — or a 500-file course folder in one go — and it converts everything to clean Markdown locally, AI-names every file, and serves the whole library to AI clients (Claude Desktop, Claude Code, Codex, Kimi, Gemini CLI) over a local MCP server (127.0.0.1 only, token-protected). The idea: your AI tools answer dramatically better when they read a clean, indexed library instead of a pile of mystery formats.

Tauri bits that might interest folks here:

• Three webviews: the main three-pane workspace, an always-on-top transparent floating dropzone (macos-private-api), and a standalone reader window
• A bundled Python venv inside the .app for document conversion, so users install nothing
• Rust-side pipeline manager for the conversion/AI jobs, with bounded concurrency so a 500-file drop doesn't stampede
• React 19 frontend; signed & notarized dmg

Tauri 2 + Rust. Free during beta (macOS, Apple Silicon): https://github.com/kasingbbb/NoteCapt

Happy to go into how any of the above works — and equally happy to hear what I did wrong 😄