複数のGmailを切り替える alternatives: what you can drop

Searching for a faster way to switch between Gmail accounts assumes switching is one task. It is not. The same click on the avatar covers four jobs that have nothing in common, and each of the four has a documented replacement that is not a faster switcher. Three of those replacements are already built into Gmail at no cost. Working out which of the four is actually causing pain is most of the work, because the answer usually removes the switching rather than accelerating it.

Four jobs wearing the same click

Count what happens across a working day and the switches sort into four groups:

• Reading. Checking what arrived at the other address.
• Sending. Replying as the other address.
• Acting for someone else. Reading and answering a colleague's inbox on their behalf.
• Noticing. Finding out that anything arrived at all.

They look identical because they share one interface. They are replaced by completely different mechanisms. Reading is a question of where mail is delivered. Sending is a question of which addresses are registered on an account. Acting for someone else is a question of permission. Noticing is a question of what is visible without any action, which makes it the only one that belongs to the browser rather than to Gmail.

This is why "make switching faster" produces no good answer. There is no single lever that improves all four. Split them and the total amount of work goes down, because two of the four typically turn out not to be tasks at all. A switch made to see whether anything happened is not a task. It is a habit filling a gap that a visible window would close.

Reading: change where mail is delivered

If the only reason for opening the second account is to see what arrived, the switch is unnecessary. Delivery can be consolidated instead.

The feature most people used for this is Gmail's "Check mail from other accounts", which collects from another account over POP. That route now has a published end date.

After the first quarter of 2026, this feature no longer supports new users. Existing users can still use the feature until January 2027. Source: support.google.com

Gmailify is covered by the same change. So POP collection is not available as a new setup, and existing setups have a date attached.

Google names three replacements. Set up automatic forwarding at the other provider, so new messages arrive directly. Use the Gmail app for Android, iPhone or iPad, which reaches other accounts over a standard IMAP connection. Or run a one time import of mail and contacts in Gmail on the web, which moves history but does not keep syncing.

One number worth checking before planning around consolidation: Gmail documents a maximum of 5 email addresses that can be added to an account. Count the addresses first.

Forwarding configured at the source behaves differently from collection configured at the destination, and the difference shows up in failure. Collection pulls on a schedule, so a broken connection produces a quiet delay that nobody notices until a message turns out to be four hours old. Forwarding pushes at delivery, so a broken rule produces mail that never arrives, which is louder and gets fixed sooner. Loud failure is the better property for anything carrying work.

What consolidation drops is the reading switch entirely. What it costs is two things. Which address a message was sent to becomes harder to see in a merged inbox, which a filter and a label fix. And spam filtering now happens twice, once at the forwarding source and once at the destination, which is harder to trace because half of it happens somewhere the destination account cannot see.

Sending: register the address instead

If the switch exists only to reply from the other address, Gmail supports adding send-as addresses under the account's name settings. Once added, the from address becomes a choice on every reply, and there is a setting to answer from whichever address received the message.

That drops the sending switch. The cost is a new category of mistake: a reply going out from the wrong address, visible to the recipient. Setting replies to default to the receiving address removes almost all of it, because the failure then requires a deliberate change rather than a forgotten one.

There is a distinction that causes real confusion here. Being able to send as an address is not the same as being able to read it. Registering send-as alone means replies to those messages land in the original mailbox, wherever that is, and nothing about that mailbox becomes visible. Setting up sending without also settling delivery produces the specific and unpleasant situation of sent messages whose answers never appear.

Before adding a work address to a personal Gmail account, check the employer's policy. Technical possibility and permission are separate questions, and this one is usually written down somewhere.

Adding a send-as address also requires confirming a link delivered to that address, so the ability to send is gated on holding the mailbox rather than on typing the address correctly. That confirmation step is a small piece of protection worth understanding: it means nobody can register an address they do not control, and it means an address can be removed from the sending list without touching the mailbox itself.

Acting for someone else: delegation instead of a password

When the second inbox belongs to another person, sharing the password is the common approach and delegation is the documented one.

A delegate is someone who can read, send, and delete messages for you. When they send a message from your account, their email address displays. Source: support.google.com

The published limits are up to 10 delegates on a personal gmail.com account and up to 1,000 on a work or school account, with typical use supporting 40 delegates accessing an account at the same time. A delegate cannot change the password and cannot use chat. Several integrated features are also outside the delegate's reach, including Google Account settings, Drive attachments, client side encryption, Gemini in Gmail and Meet in Gmail.

What delegation drops is credential sharing and the switching that goes with it. What it costs is the feature gaps above, plus the fact that a delegate's own address appears on messages they send. That second point reads as a limitation and functions as the opposite. With a shared password, nothing in the service records who sent a message or who deleted one, and the record only becomes valuable on the day somebody needs it urgently.

Noticing: a window that is already open

What remains is the switch made to check. No task attached, just a look. Counted honestly, this is the most frequent of the four.

It disappears when each account has somewhere permanently visible, because nothing needs to be switched to in order to be seen. This is the one job that belongs to the browser, and the relevant detail is how sessions are held. Google documents that when several accounts are signed in at once, settings from the default account may apply, naming Web and App Activity and Ads Personalization, and that the default is in most cases whichever account signed in first.

So keeping several accounts signed in simultaneously leaves behaviour dependent on sign-in order. Avoiding that means separate cookie stores rather than one store holding several sessions. Two shapes do this: separate browser profiles, or a tool that gives each app its own persistent window with its own session, which keeps two accounts of one service open side by side rather than stacked. How that boundary is drawn is covered in Workspaces.

This drops the check-in switch and the sign-in order dependency. It costs screen space and memory, since windows that stay open stay resident.

The memory cost is worth sizing rather than guessing, because it decides how many windows are reasonable. Each persistent window is a separate renderer process holding its own page, so the figure scales with how heavy the services are rather than with how many there are. A mail client and a chat client kept open cost far more than four documentation sites. The practical approach is to promote services to permanent windows one at a time, in order of how often they get checked, and stop at the point where the machine notices. For most setups that ceiling arrives later than expected, and the first two or three promotions remove the majority of the switching.

What should not be dropped

Two things survive the consolidation exercise, and both are worth naming because the momentum of tidying tends to take them out.

The first is the account boundary itself. Data in an employer-issued or client-issued account belongs to the issuer, and forwarding it into a personal address is prohibited by policy in most organisations. Crossing that line for convenience creates a problem that only surfaces at offboarding, when the data cannot be recovered from where it now sits.

The second is independence of two factor authentication. Holding codes for several accounts on one device is fine. Building a setup where losing that device locks every account at once is not. Either store backup codes somewhere separate, or register a second device for verification, before consolidating anything else.

What each replacement costs

Job	Replacement	What it drops	What it costs
Reading	Forwarding at the source, or the mobile app over IMAP	The checking switch	Harder to see which address received a message, spam filtering split across two places
Sending	Send-as address on the account	The reply switch	Room for sending from the wrong address
Acting for someone else	Delegation	Shared credentials	Feature gaps for the delegate, delegate's address shown on sent mail
Noticing	A permanent window per account	The habitual check	Screen space and memory

Three of the four are Gmail features and cost nothing. Only the fourth involves choosing software, and the qualifying question is narrow: does the product keep two accounts of the same service signed in at the same time, without one displacing the other.

That question comes before price. Free tiers in this category are limited along different axes, some counting apps, some counting bundles, some counting how many instances of one app can run. For multiple accounts of one service, only the last axis decides anything. Where each product draws that line is set out in Compared with Wavebox and Compared with Ferdium, and the services that can be run as their own window are listed under Supported apps.

What to change first

Spend ten minutes counting the switches by job. Most people find one job doing the damage and three that were never really a problem. Apply the replacement for that one job only. If it turns out to be the fourth, the thing to evaluate is per-window session separation, which is what SpaceDeck is built around on macOS.