Tool Friday #20 — Substack: An 8/10 I Don't Love

I’ve shipped eight issues of Inference, my newsletter on AI tooling, since late February. All of them on Substack.

If I score it honestly on what it delivers, it’s an 8 out of 10. It hosts the publication, delivers the email, keeps the archive, manages subscribers, and has cost me zero euros and — this is the part that matters — zero maintenance hours since issue #1.

And I don’t love it. Eight issues in, not one feature has made me say wow. Not the editor, not the stats page, not the publishing flow. Everything is fine. Nothing is delightful.

Most reviews are either fan letters or takedowns. This one is neither, because the most honest thing I can say about Substack is the most boring sentence in this series: it works, it asks nothing, and it never once impressed me. It took me eight issues to understand that’s a feature.

TL;DR: Substack gives a newsletter everything it needs — hosting, delivery, archive, subscriber management — for free, with literally zero infrastructure work. Publishing is free; they take 10% plus card fees only if you charge for subscriptions. The trade: no API (publishing is copy-paste by hand), flat default social cards, an editor that’s merely fine, and not a single moment of delight anywhere in the product. 8/10 on substance. If you want to write instead of running infrastructure, start here. If you want to automate or control everything, this is the one tool that won’t let you.

What it actually does

For a writer, Substack is four products pretending to be one button: web hosting for every issue, an email delivery system that actually lands in inboxes, a permanent archive, and a subscriber database with payments bolted on for whenever you want them.

The pricing model is the cleanest part: publishing is free, sending is free, and Substack only makes money when you do — 10% plus card fees on paid subscriptions. Inference is free, so my bill after eight issues is exactly zero. Not “free tier with limits I’ll hit.” Zero.

What that bought me in practice: since February I have not thought about email deliverability once. Not once. Anyone who has ever run their own sending domain knows what those words are worth.

The “fine” tax

Here’s my actual complaint, and it’s not a bug report: nothing in Substack is wow.

The editor is fine — I draft in markdown elsewhere and paste in. The stats are fine. The default social cards are flat enough that I build my own OG image by hand for every issue, because the native ones don’t do the work. The publishing flow is fine.

And there’s no API. Every other piece of my stack is scripted — builds, deploys, CRM, scraping. Substack is the one island where I’m a human copy-pasting text into a web editor like it’s 2015. Issue after issue, it’s the only fully manual publishing step left in my week.

That’s the trade in one image: the tool that asks nothing of me also gives me no way in.

What I’d want a friend to tell me before starting

There is no automation. Accept it or leave. No publish API, no programmatic anything that matters. If your identity is “I automate everything,” Substack will be the exception that mildly irritates you forever. Budget the copy-paste and move on.

Treat it as infrastructure, not as a growth channel. Substack markets a network — recommendations, discovery, the app. Plan as if none of that exists. It’s email hosting that works. If discovery sends you readers, that’s a bonus, not a strategy.

Make your own social cards. The default link previews are forgettable. If you care how your issues look when shared, you’ll end up hand-building images like I do. It’s the one recurring chore the platform creates instead of removing.

Your list is yours. Subscribers export to CSV whenever you want. The lock-in is lower than it looks, which ironically is one of the better reasons to relax and stay.

Who should bother

Start here if: you want to write a newsletter, not operate one. Zero setup, zero bill, zero maintenance — every hour goes into the writing. For getting from “I should write a newsletter” to issue #1 shipped, I don’t know anything faster.

Skip it if: you need API access, design control, or automation hooks — or your newsletter is core product infrastructure that has to plug into the rest of your system. You’ll fight the platform’s indifference and lose.

The verdict

Score: 8/10 — the highest score I’ll ever give a tool I don’t love.

Eight issues, zero infrastructure hours, zero euros. That’s the scoreboard, and the scoreboard doesn’t care that the product never sparks joy.

Would I pay for it? It’s free, so I don’t have to — and honestly, that’s load-bearing in the score. Would I move? Migrating buys me delight and costs me a working system. Eight issues in, that math doesn’t close.

The lesson travels: a tool you never think about is doing its job. Wow is optional. Shipped is not. Substack is at substack.com — set your expectations to “boring,” and it will quietly exceed them.

Tool Friday is a weekly series where I review one tool I actually use. This week: the platform behind my own newsletter — and the case for an 8/10 you’ll never fall in love with.