Honest comparisons with the email clients we respect.

How we approach a comparison

Comparison pages on the open web tend to be one of two things: an affiliate funnel that flatters the writer’s product into a strict superset of every other product, or a thin matrix of green ticks and red crosses that hides the judgement calls behind a checkbox. We’d like to write neither.

Each page under /vs/ is written in the first person. Where the other client is better, we say so plainly. Where Epistles is the right pick, we say that too, and we say for whom, because no email client is right for everyone. We don’t claim parity on a feature we haven’t shipped. The rough rule we follow: if a Newton, Spark, or Superhuman user read our page about their current client and came away thinking we were unfair to it, we’d rewrite the page.

One sentence of positioning, said once, then we let the comparisons speak for themselves: Epistles exists because no other client unified all five of the providers we use, on all five of the platforms we use, with local-first storage and a zero-knowledge vault for the credentials that move between devices. If that sentence describes a problem you have, the pages below are for you.

Comparisons available today

Two pages are written and live. They’re the comparisons readers ask for most often, and they set the template for the rest.

Epistles vs Newton Mail

Newton was the unified, keyboard-aware, multi-account client a generation of power users loved without qualification, and it shut down twice. Our page names the four things Newton got right, the parts Epistles already inherits, the parts we still owe its old users (snooze, send-later, richer sender profiles), and where we deliberately go further (local-first, zero-knowledge sync, Linux as a first-class build). It’s a love letter, not a takedown.

Epistles vs Mimestream

Mimestream is the most polished Gmail-API-native Mac client we’ve used, and if you live entirely in Gmail and entirely on macOS, it’s an excellent choice. The page is honest about that. Where Epistles answers a different question is the multi-provider matrix (Gmail, Microsoft 365, Fastmail, ProtonMail, IMAP) and the cross-platform one (macOS, Linux, iOS, Android shipping today). Same Gmail-API protocol underneath; different scope.

Comparisons coming soon

The pages below are scheduled but not yet written. Until each one ships, here is the honest two- or three-sentence framing we’d give you over email if you wrote in to ask.

Epistles vs Proton Mail (coming)

This one resolves a category-confusion question we get often: Proton Mail is an email provider, the servers and the encryption-at-rest and the Swiss jurisdiction. Epistles is an email client that supports Proton accounts alongside your other providers, with the same zero-knowledge decryption Proton Bridge gives you, in a window that also holds your Gmail and your work Microsoft 365. We replace neither Proton’s service nor Proton’s own app, we let you keep them and add four other inboxes next to them.

Epistles vs Spark (coming)

Spark is a multi-account client too, and a lovely one, with team features (shared drafts, comments) we don’t have. The architectural difference is that Spark routes mail content through Readdle’s servers to power its cross-device sync, and Epistles is local-first with a zero-knowledge vault that holds only credentials, never mail. If team collaboration is your priority, Spark is the right pick. If server-side mail content gives you pause, ours is.

Epistles vs HEY (coming)

HEY is a different thesis entirely: a workflow rebuild on a fresh domain, with the Imbox / Feed / Paper Trail vocabulary and the Screener up front. People who love that vocabulary love it deeply, and we’d send you there if it sounded right. Epistles is the other answer, the one for people who want their existing Gmail, Fastmail, and Microsoft 365 accounts as they are, in a client that doesn’t ask them to learn new nouns for an inbox.

Epistles vs Superhuman (coming)

Superhuman is $30 a month, with a real human onboarding call, an AI-summarisation surface, and a famously fast keyboard model. Epistles is $35 a year, with no AI, no concierge, and a keyboard model in the same tradition. If the AI features and the white-glove onboarding are the draw, Superhuman is the honest answer. If the price gap matters, or if you need Linux or self-hosted IMAP in the same client, this is where we land.

Epistles vs Airmail (coming)

Airmail’s Mac integration runs deep, and its plugin system has unusual third-party hooks we don’t replicate. Update cadence has been slow since the 2020 subscription pivot, and that’s a fair question to ask. Epistles is in active development with a public changelog, ships on Linux and Android in addition to Mac and iOS, and stores everything locally in SQLite. If Airmail’s plugins are what holds you there, stay; if active development matters more, we’d like a turn.

Epistles vs Apple Mail (coming)

Apple Mail is built into the operating system, free, and well-tuned for iCloud and Gmail accounts. For a Mac-and-iPhone user with one or two mailboxes, it’s genuinely fine and we’d not argue. Where Epistles fits is multi-provider depth on the providers Apple Mail flattens to IMAP (Microsoft 365 via MS Graph, Fastmail via JMAP, ProtonMail via its own protocol), and platforms beyond the Apple side (Linux, Android).

Epistles vs Thunderbird (coming)

Thunderbird is the closest in spirit on Linux: open source, cross-platform, long-lived, multi-account. The differences are mobile (Thunderbird’s Android app, formerly K-9, is shipping; iOS is in build, but our iOS app is live), the native-protocol matrix (we use Gmail API, MS Graph, JMAP, Proton with OpenPGP rather than IMAP for everything), and the Cloud Vault for cross-device credential sync. Our source isn’t open yet; theirs has been for two decades. We’ll have a fairer page once ours is.

Epistles vs Spike (coming)

Spike turns email into a chat-style timeline, with messages collapsed into conversational bubbles and the headers tucked away. Some people love it and their love is real. Epistles is the opposite design choice: traditional editorial email, with subjects, signatures, threads, and a composition surface that treats a long email as a long email. If chat-mail is what you want, Spike is right; if you write the kind of mail that needs to look like mail, we are.

Clients we considered and didn’t list

A few products you might expect to see here don’t have a page on the roadmap, for plain reasons.

• Outlook (the client itself, on Mac and Windows). The audience for “should I switch from the Outlook app to a third-party client” is largely solved by the existing Microsoft and Mimestream comparisons; a dedicated page would mostly recapitulate them.
• Inbox by Google. Google retired it in 2019. People still search for replacements, and our Newton page covers most of that audience already; a dedicated Inbox page would be addressing a six-year-old absence with thinner traffic than it earns.
• Mailspring. An honest case: Mailspring is open source, cross-platform, and we read its source while building parts of our own feature set. The audience overlap is real but small, and we’d rather cite it in the engineering blog than pretend our products are in head-to-head competition.
• Edison Mail. Edison’s 2020 disclosure that they monetised user inbox data through their analytics arm is a public matter of record; we don’t want a page whose argument writes itself.
• Polymail. Update cadence has been quiet for years, and we don’t want to litigate that without solid signal that the team is still active. If the project picks back up, we’ll add a page.

If we’re missing one

If the client you’re weighing isn’t on this page, write to us at [email protected] and tell us which one and why. The roadmap above is informed; it isn’t fixed. A few of the comparisons we ended up writing first were prompted by readers asking for them.

Or, if you’ve read enough, the rest of the site is back at epistles.com. Early access is open on macOS, Linux, iOS, and Android, fifteen-day trial, no card up front, three accounts on the free tier afterwards.