Best desktop mail clients, and when webmail as an app wins

The search for the best desktop mail clients usually starts with a specific irritation rather than a shopping list. Mail opens in a browser tab that keeps drifting to the far right of the tab strip. A message arrives and the notification comes from the browser, not from anything that looks like a mail app. Command and W closes a half written reply. None of that is a mail problem. It is a window problem, and the two have different solutions.

A desktop client solves it by taking mail out of the browser entirely, with its own database, its own notifications and its own place in the Dock. That is the right answer for a lot of setups, and the candidates below are the ones still actively shipping on macOS. It is not the only answer, though, and the second half of this article covers the cases where installing a client trades one set of problems for another.

What a desktop client actually changes

Four things change the moment mail lives in an installed application, and they are worth naming precisely because only one of them is about features.

A local copy of the archive. A client downloads messages to disk. Search runs against that local store, so it returns results with the network off and it is not limited to whatever search syntax the provider exposes. For an archive that goes back a decade, this is the single biggest practical difference.

Notifications come from the mail app rather than from a browser, so they carry the right icon, respect a Focus schedule aimed at that one app, and show a badge count on a Dock icon. Small, but it is the difference between noticing a message and not.

Accounts sit side by side. Several providers in one unified inbox, with one keyboard shortcut set across all of them, removes the habit of keeping three webmail tabs open and checking each one in turn. Providers do not cooperate on this. A client is the only place it happens.

The window is a mail window. It does not have a tab strip above it, an address bar, or a keystroke that closes a draft because the browser thought the shortcut meant something else. Every one of the clients below gets this right, and so does any properly installed web app, which is the point the comparison later turns on.

The clients still worth shortlisting

Apple Mail

Already installed, no account to create, no price. It handles IMAP, Exchange and Gmail, sits inside the system share sheet, and works with Contacts, Calendar and iCloud Drive without configuration. Apple keeps a full Mail User Guide for it. The weakness is provider specific features: Gmail labels, filters and search operators map onto Mail's folders and rules imperfectly, and heavy Gmail users tend to feel that friction within a week.

Thunderbird

Free and open source, funded by donations, with builds for macOS, Windows and Linux plus an Android app. It handles mail, calendar and contacts in one application, offers a unified inbox across accounts, and has a tag system that survives serious volume. It is also the most customisable option here by a wide margin, which is why it keeps appearing in recommendations from people running six accounts across four providers. The interface is dense by default, and that is the honest trade.

Mimestream

A Mac only client built specifically around the Gmail API, which means Gmail labels, filters, categories and search behave as Gmail defines them rather than being translated into IMAP folders. Pricing is a subscription: $4.99 per month or $49.99 per year, with a fully functional 14 day trial that does not ask for a card. A licence covers up to 5 devices and can be shared with immediate family in the same household. It is sold directly rather than through the App Store, and the direct build is sandboxed, notarised and signed with a Developer ID. If every account is a Google account, this is the shortlist entry that fits best. If any account is not, it does not apply at all.

Spark

Free tier at $0 covering unlimited email accounts, a unified inbox, smart notifications and a calendar. Paid tiers add collaboration and assistant features: Plus at $10 per user per month, or $8.25 per user per month billed annually at $99 per year; Pro at $20 per user per month, or $16.58 billed annually at $199 per year; Enterprise by arrangement. There is a Windows build as well as Mac and mobile. The paid features are aimed at shared inboxes and client facing teams, so a solo user rarely needs to leave the free tier.

eM Client

Windows and Mac, with mail, calendar, contacts, notes and chat in one window, plus PGP and S/MIME encryption. The free licence requires registration, is limited to personal non-commercial use and allows 2 email accounts. Personal is $39.95 for a one year subscription covering up to 3 devices, or $59.95 as a one time payment with lifetime upgrades. Business is $49.95 per device per year, or $79.95 one time. There is a 30 day trial and a 30 day money back guarantee, and the Android and iOS apps need no licence.

Side by side

Client	Platforms	Price today	Model
Apple Mail	macOS, iOS	Included with the system	None
Thunderbird	macOS, Windows, Linux, Android	Free	Donations
Mimestream	macOS only, Gmail only	$4.99/mo or $49.99/yr	Subscription
Spark	macOS, Windows, iOS, Android	$0, or $10 and $20 per user/mo	Freemium
eM Client	macOS, Windows, iOS, Android	Free for personal use, $59.95 one time	Licence

The table answers the budget question and settles almost nothing else. Two of these are free, so cost is not the deciding factor for most people. Provider fit is. Mimestream is the best Gmail client on the list and cannot open a Fastmail account. Thunderbird will connect to anything and will not give Gmail labels first class treatment. That axis, not price, is where the decision actually gets made.

Where webmail still wins

There is a reason the switch does not stick for everyone, and it is not inertia.

Provider features ship to the web first. Gmail's search operators, label automation, scheduled send behaviour and every feature Google adds next quarter exist in the web interface on day one. A client gets them when its developer implements them, if the API exposes them at all. Anyone who lives inside Gmail's filters and labels is choosing between a better window and a worse Gmail.

Some things never leave the web at all. Workspace admin settings, Google Chat, Outlook's web add-ins and the full Exchange calendar surface are examples. If mail and a chat tool and a calendar all live behind one sign in, a desktop client covers one of the three and the browser still has to stay open for the rest.

Setup is not free either. Multiple accounts, app passwords, OAuth consent screens, signature migration, folder mapping and a first sync that can take hours on a large archive. That is real work, and it repeats on every new machine.

And there is the sync question. A client holds its own copy of state, which is the point, but it also means a second place for that state to be wrong. Read status, drafts saved in two places, a filter that ran server side after the client had already cached the message. None of this is common, and all of it is annoying when it happens.

Webmail in its own window

The middle path is to leave mail where the provider maintains it and fix the window instead. Every practical complaint in the opening paragraph is a window complaint, and macOS has two routes to solving it without installing a mail client at all.

In Safari, open the site, click the share control in the toolbar and choose Add to Dock. The result is an icon in the Dock and in Spotlight, an existing sign in that carries over, a simplified toolbar without the tab strip and address bar, and an app that can post notifications like any other. In Chrome, the equivalent sits in the three dot menu under Cast, save and share, as Install page as app, and Google's help pages cover managing the result.

Both routes give the Dock icon, the Command and Tab entry and the reclaimed vertical space. What separates them, and what separates both from a purpose built site to app tool, is what happens underneath. Safari web apps do not run browser extensions and do not isolate one app's cookies from another's, which rules out both a password manager extension and two accounts with the same provider signed in at once. A Chromium based wrapper with per app profiles handles both, which is the specific reason people with a work Google account and a personal one end up there. The Features page lists which parts of that window can be adjusted per app, and the Supported services list shows which mail providers already have presets.

	Browser tab	Installed web app	Desktop client
Command and W closes	The draft	The app window	The app window
Offline archive	No	No	Yes
Provider's newest features	Immediately	Immediately	When implemented
Setup time	None	Under a minute	An hour, plus first sync
Multiple accounts in one place	No	One per app	Yes

Deciding in one afternoon

The question that settles it is not which client is best. It is whether the offline archive and the unified inbox matter enough to pay the setup cost and accept the lag on provider features.

A test that takes an afternoon rather than a fortnight: install the web app route first, because it costs under a minute and requires no decision. Use it for a week. If the only remaining complaints are about the window, the work is finished. If the complaints that survive are about search across ten years of mail, or about checking four providers in turn, then a client is warranted, and by then the shortlist has narrowed on its own. All Google accounts points to Mimestream. Mixed providers or an unusual setup points to Thunderbird. Encryption and a one time payment point to eM Client.

Running both is also normal, and worth saying plainly. A client for the archive and the serious triage, a web app window for the provider features that only exist on the web, each with its own icon. Nothing about the two choices is exclusive.

What to change first

Start with the window, because it is the cheapest change and it removes the complaints that a new mail client would also have removed. Give mail an icon, a Dock position and a window with nothing else in it, then judge after a week whether a full client is still needed. A tool such as Kagemusha handles the per app profile and extension side of that, and the Pricing page covers what the unlimited tier costs.