Obsidian Publish Site Downloader

Download any Obsidian Publish site locally, even if it's password-protected.

Needed a way to back up and download entire Obsidian Publish sites. Since there's no official "export" feature for viewers, doing this manually is incredibly tedious.

This script handles the heavy lifting. It authenticates just like the official web app, fetches the site's internal manifest, and downloads everything in parallel. It even handles Windows filename restrictions so your files won't get messed up.

Highlights

• Bypasses passwords: Fully supports password-protected sites using Obsidian's native authentication handshake.
• Fast & Concurrent: Downloads multiple files at once (10 threads by default).
• Resumable: If your connection drops, it skips files you already have.
• Smart caching: Type your password once; it caches your token for future runs.
• Resilient: Automatically retries failed requests with exponential backoff.

Getting Started

The easiest way to run this is with uv, but standard Python and pip work perfectly fine too.

First, install the dependencies (requests and tqdm):

# With uv (recommended)
uv pip install requests tqdm

# Or standard pip
pip install requests tqdm

Usage

uv run download.py [options] <URL> <FOLDER>

Downloading a public site

Just point it at the URL and give it a destination folder. It will create the folder if it doesn't exist.

uv run download.py https://help.obsidian.md/ vault

Downloading a password-protected site

Add the --password flag to securely enter the site's password.

uv run download.py https://my-private-site.com/ vault --password

Note: Your token gets saved as .obsidian-publish-token in your destination folder. On future runs, it uses this cached token automatically so you don't have to keep typing your password.

Resuming a large download

If you have a massive site or your internet cuts out, you can run the same command with --resume to skip files you've already downloaded. Feel free to crank up the --workers if you want it to go faster!

uv run download.py https://example.com/ vault --resume --workers 20

Options

Here are the different flags you can pass to the script:

Flag	Description	Default
-p, --password	Prompt for the site's password	Off
-w, --workers	Number of concurrent downloads	10
-r, --resume	Skip files that already exist on disk	Off
--max-retries	How many times to retry a failed file	5
--retry-delay	Base seconds to wait before retrying	3
--no-sanitize	Don't rename invalid Windows characters	Off