Substack PDF Archiver

Bulk-saves a Substack publication's posts as pixel-perfect PDFs, named YYYY-MM-DD – Title.pdf so the folder sorts chronologically. For personal, offline archival of publications you have a paid subscription to.

⚠️ Personal use only. This tool is for making offline backups of content you legitimately pay for. Do not use it to redistribute, share, or republish paywalled posts, and respect each publication's terms of service. You are responsible for how you use it.

Setup

cd ~/Downloads/CLAUDE/Code/substack-pdf-archiver
npm install
npx playwright install chromium

Dashboard (easiest)

npm run dashboard

Open http://localhost:4321. From there you can:

1. Connect Substack account — opens a browser, you sign in, click “I'm signed in”. One login covers every publication you subscribe to.
2. Paste a publication URL (full URL or just the slug).
3. Download all posts (checkbox, on by default) — or uncheck it and pick a From / To date range.
4. Hit Download and watch live progress; Stop cancels mid-run.

PDFs land in pdfs/<publication-host>/. Already-downloaded posts are skipped.

The command-line workflow below does the same thing if you prefer a terminal.

---

1. Log in (one time, per account)

npm run login

A browser window opens at Substack's sign-in page. Log into your account, then press Enter in the terminal. The session is saved to storageState.json.

One login covers every publication you subscribe to — Substack uses a single account sign-in, so you don't log in separately per publication.

2. Archive

# Default publication (set in config.js)
npm run archive

# Any publication — full URL, or just the slug
node archive.js https://foo.substack.com
node archive.js foo                    # shorthand for foo.substack.com

PDFs land in pdfs/<publication-host>/, so each publication gets its own folder.

Incremental downloads (--since / --until)

To grab only new posts since you last ran it — no re-scanning the old ones:

node archive.js foo --since 2026-05-31
node archive.js foo --since 2026-05-31 --until 2026-06-15

Posts come newest-first, so --since stops paging the API as soon as it crosses below the date — fast even on huge archives. Dates are YYYY-MM-DD.

Even without --since, existing PDFs are skipped, so re-runs never re-render. --since is just faster because it doesn't even visit the old posts.

Other behavior

• Pixel-perfect: screen styling + backgrounds, scrolls to load lazy images.
• Throttled: randomized 3–5s between posts.
• Resumable: stop/restart anytime; re-run to retry failures.

Config

config.js holds defaults: the fallback publication, output base dir, throttle range, and API page size.

Notes

• Personal use only — don't redistribute paid content.
• If PDFs start showing paywalls, your session expired — re-run npm run login.