Substack Scraper — Posts, Likes, Paywall Status & Cadence

What does this newsletter actually publish, how often, and how much of it is behind the paywall?

Every post a Substack publication lists, newest first, with likes, comments, restacks, word count, tags, author and whether it sits behind the paywall. Or flip the switch and get one summary row per newsletter — publishing cadence, paid share, median engagement — for a whole list of them in one run.

Works on .substack.com addresses and on custom domains alike. No API key. No login. No proxies.

---

Two modes

Posts — one row per post. The default. Point it at a publication and get its archive as a table: title, subtitle, date, type (newsletter, podcast, thread), paywall status, likes, comments, restacks, word count, tags, authors, public excerpt, link. Turn on the body option and free posts come with their full text.

Summary — one row per publication. Give it twenty newsletters and get twenty rows you can sort: posts per week, share of paywalled posts, median likes and comments, median length, days since the last post, the tags it keeps returning to, and the subscriber band the publication states about itself.

Measured 11 September 2026 on three publications, summary mode, newest 100 posts each:

Publication	Posts / week	Paid posts	Median likes	Median comments	Median words
Lenny's Newsletter	0.78	75%	335	10	2,955
The Pragmatic Engineer	2.41	62%	156.5	5	3,935
Noahpinion	3.73	26%	430	82.5	2,884

Three very different businesses, visible in one table.

---

What you get

Posts mode

The post — postTitle, subtitle, postUrl, postType, section, excerpt, coverImageUrl, postId

When — publishedAt, daysSincePublished

Who can read it — paywalled (true/false) and the raw audience value Substack uses (everyone, only_paid, founding)

Engagement — likes, comments, restacks

Content — wordCount, podcastDurationMin, tags, authors, and body as plain text when you ask for it (free posts only — see below), with hasFullText telling you which rows have one

Where it came from — publication (the domain it actually lives on) and publicationName

Summary mode

publication, publicationName, publicationAuthor, publicationUrl, subscribersBand, postsRead, postsLast30d, postsLast90d, postsPerWeek90d, paidPostsPercent, podcastPostsPercent, medianLikes, medianComments, medianWordCount, latestPostAt, daysSinceLastPost, oldestPostReadAt, topTags, authors, status

---

Typical uses

• Size up a category of newsletters. Summary mode on a list of competitors: who publishes daily, who has gone quiet, who charges for most of their writing, who gets the comments.
• Track one publication over time. Posts mode with a 7- or 30-day window on a schedule, and you have a dated log of every post and how it landed.
• Find what works. Sort a year of posts by likes or restacks and read the titles. Filter to free posts and pull the bodies for a content brief.
• Check a sponsorship or partnership target. Cadence, paid share and the publication's own subscriber band, before the call.

---

Input

Field	What it does
publicationUrls	Newsletters to read. https://www.lennysnewsletter.com, noahpinion.substack.com, or just noahpinion. Custom domains work; a .substack.com address that has moved to a custom domain is followed there.
mode	posts (one row per post, default) or summary (one row per publication).
daysBack	0 for no date filter; 30 for the last month. Applies in both modes.
maxPostsPerPublication	Cap per publication, newest first. Default 100; up to 5,000. In summary mode this is what the figures are computed from.
includeBody	Add the full text of free posts as plain text. One extra request per post. Posts mode only.

Nothing else is needed. Substack serves the archive of every public publication without a login, and this actor reads exactly that.

---

What it costs

$3.00 per 1,000 posts in posts mode, $10.00 per 1,000 publications in summary mode, plus $0.005 per run. Apify's Starter, Scale and Business plans get 10%, 20% and 30% off.

You pay for rows delivered, not for requests. A private publication, an address that is not on Substack, or a filter that matches nothing costs nothing. Fetching post bodies costs nothing extra.

For scale: one publication's whole archive is usually 100–600 posts, so a typical run is a few cents. Summary mode on 50 newsletters is 50 cents.

---

How it works

Substack publishes every public newsletter's archive as JSON at /api/v1/archive — the same data the publication's own archive page is rendered from — and each post at /api/v1/posts/<slug>. This actor reads those two endpoints and nothing else: no HTML scraping of post pages, no browser, no proxies. That is why it is cheap and why it does not break when a publication changes its theme.

The archive is walked newest-first in pages of 50 until the date filter, the cap or the end of the archive is reached. Where a .substack.com address redirects to a custom domain, the redirect is followed and the row reports the domain the publication actually lives on.

The publication's name and its subscriber band come from the sentence Substack writes into every publication's page description — "a Substack publication with tens of thousands of subscribers". That band is the only size figure Substack makes public, and it is reported in the publication's own words rather than turned into a number.

---

Honest limits

• Paywalled posts never come with a body. Substack's post endpoint will hand most of a paywalled post's text to anyone who asks. This actor does not ask. Paywalled rows carry the title, date, engagement and the public excerpt, hasFullText is false, and body is empty. If you need the text of a paid post, subscribe to the publication.
• Likes, comments and restacks are the counts at the moment of the run. They keep changing on Substack; a scheduled run gives you the curve.
• subscribersBand is a band, not a number — "thousands", "tens of thousands", "hundreds of thousands", "millions" — because that is all Substack publishes. It is blank when the publication has not stated one.
• Cadence figures in summary mode are computed from the posts read. If the cap is reached before the 90-day window is covered, postsLast30d, postsLast90d and postsPerWeek90d are left empty rather than reported as a floor, and status says so. Raise the cap to fill them in.
• Private publications (invite-only) answer with HTTP 403 and are reported as such, with no rows. Publications that have left Substack for another platform answer 404 and are reported the same way.
• tags is empty for publications that do not tag their posts, and section is empty for those without sections. That is the publication's choice, not missing data.
• The whole archive can be long. Daily publications have thousands of posts. The default cap of 100 keeps a first run quick; raise it when you want everything.

---

Scheduling

Newsletters move weekly. Schedule posts mode with daysBack: 7 on your list and every run adds the week's posts with their engagement at that moment; schedule summary mode monthly and you have a dated record of each publication's cadence and paid share over time.

---

Legal

This actor reads only data that Substack serves publicly for each publication, at addresses it serves to anyone. It collects no personal data beyond author names the publication itself prints on its posts, requires no login, bypasses no access control, and never retrieves the text of paywalled posts. You remain responsible for how you use the exported data.

---

Questions, or a publication that behaves oddly? Open an issue on the actor — issues are read and answered.