The Case Against Subscription Email: Software You Actually Own

Email is the most basic tool you have. It predates the web. It’s an open standard you can reach with a dozen free apps. So there is something faintly absurd about paying a monthly fee, indefinitely, for a program whose job is to show you your own mail.

This is not an argument that subscriptions are evil. Plenty of software should be a subscription. The argument is narrower and, I think, harder to dodge: a desktop email client, the thing that renders messages already sitting on your provider’s servers, is a poor fit for renting. Yet the industry has quietly moved you onto rent anyway, and it is counting on you not noticing. C+R Research asked people to guess their monthly subscription spend and then itemise it: the average guess was $86, the actual figure was $219. Recurring charges are designed to disappear from view. An email client is a strange thing to let disappear into that pile.

When a subscription is legitimately fair

Credit where it’s due. Some email subscriptions fund real, ongoing costs:

• Server-side features. If a client runs its own infrastructure on your behalf (cloud sync, server-side AI, shared team state, push relays), that costs money every month it’s switched on. Renting it is reasonable.
• A genuine platform tax. A client built on the Gmail API inherits a recurring, Google-imposed cost: an annual third-party security assessment it must pass and renew every year to keep access, plus exposure to Google’s quotas and rule changes. (We unpack this in IMAP vs the Gmail API.) Mimestream is upfront that this is part of why it charges $49.99 a year, and that’s honest of them.
• Constant, heavy R&D that genuinely ships new capability every month rather than maintenance.

If you’re paying for one of those, fair enough. You’re renting something that really does cost the developer money every month.

But a lot of subscription email clients aren’t doing any of that. They’re a desktop app reading IMAP mail, the same job a one-time purchase did a decade ago. Nothing got more expensive to run. The pricing strategy changed, because recurring revenue is wonderful for a software business and merely fine for the person paying.

The real cost: do the maths

A monthly price hides its own size. Stretch it over the time you’ll actually use a client (five years is conservative for a tool you live in daily) and the gap gets stark. Approximate five-year totals, using each vendor’s cheapest sensible individual plan on annual billing, checked June 2026:

A couple of caveats to keep this honest. HEY bundles actual email hosting and a @hey.com address, so it’s funding real servers, not just an app. Spark and Superhuman ship genuine server-side AI features. And prices drift; these are the figures at the time of writing, rounded, excluding tax.

Still, look at the spread. The recurring options don’t cost five or ten times more because they do five or ten times more. They cost more because the meter never stops. A one-time client at the top of that table does the core job the $1,500 one does: it shows you your mail, fast, on your machine. (We’ve done detailed side-by-sides against Superhuman, Spark and Mimestream if you want the feature-level view.)

The graveyard: what history says about renting your inbox

Money is the obvious half of the argument. The other half is what happens when the company behind your client loses interest, and email clients have an unusually grim track record here.

Sparrow was the darling Mac client of 2012 until Google bought it and development stopped dead. Dropbox paid roughly $100 million for Mailbox in 2013, then shut it down in February 2016 when it no longer fit the strategy. Astro lasted about a year after Slack acquired it in 2018. Google killed its own Inbox in 2019. And Newton Mail, a $50-a-year subscription client, shut down in 2018, was revived, shut down again in 2020, was revived again, and finally closed for good in July 2024. Its paying subscribers went through that cycle three times.

This keeps happening because a subscription client isn’t really a product you hold; it’s a company you depend on. When Grammarly acquired Superhuman in July 2025 (and later renamed the whole company Superhuman), the app didn’t vanish. But its roadmap now belongs to a productivity-suite strategy its subscribers never signed up for, and they have no lever beyond cancelling.

When you’ve bought software outright, an acquisition or a pivot can’t reach into your machine and switch off the copy you paid for. When you rent, it can, and historically it does.

You lose it when you stop paying

There’s a quieter version of the same problem that doesn’t need a shutdown at all. Cancel a subscription client and, in most cases, the tool you organised your working life around becomes a login screen. With a perpetual licence, stopping payment just means you stop receiving new versions. The software keeps working exactly as it did the day before.

That asymmetry is the whole game. A subscription’s leverage is that quitting costs you the product. A one-time purchase has to earn your money once, honestly, up front.

A stronger backstop: source-available software

There’s a version of ownership that goes further than a perpetual licence: software whose source is available. If the company disappears, the code doesn’t. It can be built, patched and kept running independently of any one business surviving. For a tool as load-bearing as your email client, that’s a meaningful insurance policy no closed subscription app can offer. Kanmail happens to be both source-available and a one-time purchase, which is not a coincidence; the two ideas come from the same place.

Who should pick a subscription

To stay honest: a subscription is the right call if you genuinely use the server-side features it funds. Heavy AI assistance, real-time team collaboration, shared inboxes with synced state. If those are core to your day, renting them is rational, and teams on Missive or Shortwave are usually paying for exactly that. The case against subscriptions is a case against renting a plain desktop client that does none of it.

What to look for if you’d rather own

• A perpetual licence: buy once, keep that version forever
• No usage limits or feature gates tethered to a live subscription
• Local-first, with mail on your device, so the app isn’t hostage to a service staying online
• A free, time-unlimited trial, so you prove the fit before paying anything
• Bonus: source-available, so the software can outlive the company

FAQ

Is there a genuinely good email client without a subscription? Yes. Thunderbird is free; eM Client, Mailbird and Kanmail sell one-time licences. You are not obliged to rent.

Aren’t one-time apps abandoned faster? Not inherently, and the graveyard above is mostly subscription and VC-funded apps. One-time software is built to survive on one-time revenue, and perpetual licences with paid major-version upgrades have sustained desktop software for decades. Source-available software is the most resilient of all.

Is a subscription ever worth it for email? When you actually use the server-side or AI features it pays for, yes. For a plain client that just shows your mail, you’re renting something you could own.

You already own your email. It’s an open standard on servers you can reach a hundred different ways, and the client is just a window onto it. A window shouldn’t come with a meter.

Kanmail is a one-time $49 purchase: no subscription, no usage limits, free to try for as long as you need, source-available, and your mail never leaves your machine. Buy it once and it’s yours.

Related: IMAP vs the Gmail API · How to Manage Multiple Email Accounts · Switching from a subscription client to Kanmail