gugatec/markdowndown repository overview

⁠MarkdownDown

Convert any webpage to clean Markdown, with images downloaded and embedded. Self-hosted fork of asadm/markdowndown⁠ (original live demo⁠) — telemetry removed, packaged for Docker.

Source: https://github.com/Gugatec/markdowndown⁠

⁠Quick start

docker run -d \
  --name markdowndown \
  -p 3000:3000 \
  gugatec/markdowndown:latest

Open http://localhost:3000, or POST a URL to http://localhost:3000/api/tomd.

⁠docker compose

services:
  markdowndown:
    image: gugatec/markdowndown:latest
    container_name: markdowndown
    ports:
      - "3000:3000"
    environment:
      # Optional: only needed for the "apply GPT" transform feature
      - OPENAI_API_KEY=${OPENAI_API_KEY:-}
      # Optional: use a remote Browserless instance instead of local Chromium
      - BROWSERLESS_KEY=${BROWSERLESS_KEY:-}
      # Optional: use a remote HTML-fetch API (e.g. a Cloudflare Worker) instead of local Chromium
      - HTMLFETCH_API=${HTMLFETCH_API:-}
    restart: unless-stopped

Save as composer.yaml and run docker compose -f composer.yaml up -d.

All environment variables are optional — a plain docker run with no env vars works for basic URL → Markdown conversion.

⁠Tags

⁠Features

• Convert webpage to markdown using Puppeteer⁠ and Turndown⁠
• Clean up content using Mozilla Readability⁠
• Download images, embed them in markdown, and download as a zip
• Optional GPT-3.5/4 pass for summarization, link removal, reformatting, etc.
• Also returns a clean HTML version of the webpage

⁠What's different from upstream

This fork removes the Mixpanel analytics call in the web UI (nothing phones home) and packages the app for Docker with a multi-arch build. See the GitHub repo⁠ for full source and Dockerfile.

⁠License

MIT — see LICENSE⁠. Original work © Asad Memon.