OneMoon

OneMoon is a human-in-the-loop note-to-LaTeX workspace. It takes handwritten notes, screenshots, or PDFs through upload, page review, block conversion, and final LaTeX packaging in one flow.

What It Does

• Upload a document into a workspace and review pages in a browser-based editor.
• Segment handwritten content into text, math, and figure blocks.
• Convert blocks with a configured LLM provider or keep using the built-in mock provider for local development.
• Merge the assembled output into LaTeX and download a package that includes the source plus a relative figures/ folder.

Screenshots

Core Flow

1. Create or open a workspace.
2. Upload a PDF or image document.
3. Review each page and adjust block segmentation where needed.
4. Convert blocks in Conversion mode.
5. Merge the assembled source in Merging mode and download the resulting package.

Stack

• Frontend: React, TypeScript, Vite
• Backend: FastAPI, SQLAlchemy, Python processing pipeline
• Storage: local data/ directory for uploads, rendered pages, crops, and generated artifacts

Run Locally

Start both apps from the repo root:

npm install
npm run dev

• Frontend: http://localhost:5173
• Backend: http://localhost:8000

Run the apps separately when needed:

cd apps/backend
uv sync --dev
uv run onemoon-backend

cd apps/frontend
npm install
npm run dev

Default Login

• Username: admin
• Password: onemoon

LLM Configuration

OneMoon stays usable without an external model. If no real provider is configured, the backend falls back to the mock provider so the upload-to-review workflow still runs locally.

To enable a real provider, configure apps/backend/.env or the repo-root .env with your ONEMOON_* / provider-specific LLM settings before starting the backend.

Verification

cd apps/backend && uv run pytest
cd apps/frontend && npm run build

More detailed app-specific notes live in apps/frontend/README.md and apps/backend/README.md.