remarkdown

A modern, clean Dockerized web app that converts Office documents, PDFs, web pages and YouTube links into Markdown — powered by a PHP frontend and a Python (FastAPI) backend running MarkItDown.

• Repository: relooplab/markdown_app
• This project is a fork of office_to_markdown_app.

---

✨ Features

• Drag & drop upload for Word, Excel, PowerPoint, PDF, HTML, CSV, JSON, XML, ZIP, EPUB and more
• Convert YouTube URLs to Markdown (transcripts)
• Split-view converter: inputs on the left, live result on the right
• Rendered or Source preview with one-click Copy and Download .md
• Output metrics: characters, words, lines, size
• Modern responsive UI with light & dark mode

🏗 Architecture

Browser ──▶ Nginx (proxy)
              ├── /  → PHP-FPM  → index.php (HTML/CSS/JS UI)
              └── /api/* → FastAPI (Python) → markitdown

Service	Stack	Port
nginx	nginx:1.27-alpine	8080
frontend	php:8.3-fpm-alpine	(internal 9000)
backend	python:3.11 + FastAPI	(internal 8000)

🚀 Getting started (Docker)

1. Build and start all services:

   docker compose up -d --build

2. Open your browser at http://localhost:8080

3. Override the host port if needed:

   PORT=9000 docker compose up -d --build   # → http://localhost:9000

Stop with docker compose down. Rebuild backend after changes with docker compose up -d --build backend.

🛠 Tech stack

• Frontend: PHP 8.3, vanilla JS, custom CSS (light & dark themes), marked.js for preview rendering
• Backend: Python 3.11, FastAPI, MarkItDown
• Orchestration: Docker Compose + Nginx reverse proxy

📁 Project layout

markdown_app/
├── docker-compose.yml
├── .dockerignore
├── frontend/
│   ├── nginx.conf          # routes .php → frontend, /api/* → backend
│   └── www/
│       ├── index.php       # homepage + converter UI
│       ├── style.css
│       ├── app.js
│       └── logo.svg
└── backend/
    ├── Dockerfile
    ├── requirements.txt
    └── main.py             # FastAPI: /api/convert, /api/convert-url

🔌 API

• GET /api/health — liveness probe
• POST /api/convert — multipart upload (file) → { markdown, file_name, characters, words, lines, size }
• POST /api/convert-url — JSON { "url": "https://youtube.com/..." } → same shape

📄 License

Project repository: relooplab/markdown_app — a fork of office_to_markdown_app.