Bottom Line: As a reader, Substack's mobile app is one of the best-built content clients on either store — chronological, ad-free, and genuinely pleasant to live in for hours. As a business platform wearing a social network's clothes, it's a far messier proposition, and the app increasingly wants you to notice.
The Reading Loop
The fundamental interaction here is beautifully mundane. Open app. Read top of Inbox. Scroll. That's it. There is no infinite scroll of stuff you didn't ask for wedged into the primary surface, no "suggested for you" card at position three, no ambient anxiety that you're missing something the ranking model decided to hide. The Inbox empties. Content apps that let you finish are rare enough now to feel almost subversive.
The reader itself is where the app's design philosophy shows its teeth — and its stubbornness. Text renders in a fixed serif with limited control over type size and spacing, and this is the single most consistent complaint in an otherwise adoring review corpus. Substack's position is defensible: they want every publication to look like it belongs to the same well-designed magazine, and typographic consistency is how you get there. But a reading app that won't let a reader adjust line height is prioritizing the brand's visual identity over the eyeballs actually doing the work. Kindle solved this in 2007. Pocket solved it. The refusal isn't minimalism; it's a decision that form should win, and it's the wrong call.
Offline caching, by contrast, is handled with real care. Downloaded posts and episodes persist properly, the library is easy to reach, and the app degrades gracefully on a subway platform instead of throwing spinners at you. Text-to-speech is the quiet standout feature. Substack's audience skews toward long essays — 4,000 words on monetary policy, 8,000 on a media feud — and the ability to convert any of them into audio without the writer having to record anything changes how much of your subscription you actually consume. That's not a gimmick. That's the difference between a paid subscription you read and one you cancel.
The Notes Problem
Here's the tension at the center of this app. Substack's revenue depends on subscription conversion, and subscription conversion depends on discovery, and discovery on the open web is dead. So Substack built its own: Notes, plus a recommendations graph where writers vouch for each other, plus video, plus livestreams. Strategically, this is sound. Experientially, users are pushing back — the growing prominence of Notes, video, and livestream content over plain newsletters is a recurring gripe, and it lands because the app's original pitch was no feed, no algorithm, no engagement bait. Every pixel Notes claims is a pixel that quietly renegotiates that deal.
To be fair to Substack, the separation is still real. Notes lives in its own tab. The Inbox has not been algorithmically contaminated. But the gravitational pull is visible in the notification behavior, which is the app's other genuine failure. Notification volume runs hot out of the box, and the defaults are tuned for a platform trying to grow a social network rather than for a reader who wanted their newsletters in one place. You can fix it. You shouldn't have to.
Where the Writer Experience Falls Apart
Mobile publishing is real but deliberately thin: draft, publish, read subscriber counts and open rates, reply to comments. For triage and community management, it's fine. Nobody is composing a 5,000-word investigation on a phone, and Substack sensibly doesn't pretend otherwise.
The friction is structural. Handling multiple accounts is clumsy — if you both read and publish, or run more than one publication, switching is a chore in an app that's otherwise frictionless. And the wider platform draws far harsher reviews on Trustpilot and TrustRadius than the app does on the stores: Stripe-only payouts that lock out entire regions, thin analytics that tell you open rates but little about what those readers actually do, weak third-party integrations, and minimal organic discovery for publications starting from zero. Support is the sharpest wound — AI chatbot first, slow human escalation, including on billing disputes where money is genuinely at stake. That's an acceptable design for password resets and an unacceptable one for revenue.
The honest summary: this app is a joy on the consumption side and a competent-but-limited companion on the production side, attached to a business platform whose economics work best for writers who already have an audience.