If you’re looking for an alternative to Gmail, may I introduce Port87.
I’ve been working on this email service for four years, and I just opened it to the public today. The way it works is a might different than other email services.
You still get an address like [email protected], but you don’t usually use that form. Instead, you add a label like [email protected]. This is often called subaddressing or plus addressing. With Port87, these addresses go into labels. Everything between the dash and the at is a label. When you sign up somewhere, you can give them a label, even if it doesn’t exist yet. Then it becomes a pending label that you can approve to move it in with all your other labels. This really helps with organization too.
You can also give labels meant for real people, like [email protected]. On labels meant for real people, you can enable screening that responds to anyone new with a link to prove they’re human. When they click the link, their email is delivered.
Lastly, you can give out your “bare address” ([email protected]) anywhere, because any email to it doesn’t get delivered to you. Instead, they get a response saying to email one of your other addresses, then a list of all of the addresses to your public labels. For example, I have a public label at [email protected] that’s meant for email about my open source projects. That gets included in the list in the auto reply when you email [email protected].
Oh, also, you can bring your own domain! The main benefit of your own domain is it prevents vendor lock in. If Port87 ever stops meeting your needs, you can pack up your domain and take it to another provider. It also prevents losing your address if Port87 ever shuts down.
If you can’t tell, I’m very passionate about email, and the more competition there is to Gmail and Exchange, the more they’ll be forced to actually stop trying to Embrace Extend Extinguish email.
Looks pretty great and I applaud the efforts. We need more services like this in the space! Pricing is hard but purelymail.com pricing really makes sense and works for me. You pay based on how much you use it and it lets you add unlimited domains. I’m sure it’s not a great fit for everyone but for light users the pricing is wonderful. If you come up with something similar I’d try your service.
Unfortunately, that’s way cheaper than I’d be able to go. If I were just running standard open source servers, it would be easier to hit that price point, but I had to write my own server and client software to make it work the way it does. It’s mostly that ongoing development cost that keeps me from offering lower prices. I’ve tried to stay competitive as best I can.
I’m also trying to run the entire thing without any outside investment, so I can’t subsidize the cost with VC funding, which is the only other way I’d be able to hit that price point. As of right now, I have zero outside investment, and I’m not currently planning on taking any.
From my perspective, no VC investment is a huge plus.
Really impressive to roll your own mail servers and clients! Do you use AI much to help with development?
None at all. For a few reasons:
- I’ve found the quality of AI generated code to be very low.
- It takes longer to fix bad AI code than to just write it correctly on my own.
- AI tends to heavily rely on RegEx, which I try to use only sparingly, because of the performance cost.
- It would open me up to potential copyright infringement and license violation claims, because the AI doesn’t know how to correctly adhere to copyright and open source licenses.
AFAICT, every AI has been trained on at least some of my code, and some of them will spit out my own code nearly verbatim if I ask them in a certain way, so I 100% know that the copyright issue makes it a no-go even if the quality problems are solved.
You can see for yourself if you ask them to generate an example of using a Svelte Material UI button. They’ll give you code that looks nearly identical to my examples available on https://sveltematerialui.com/
this is great, commenting so i remember to sign up tomorrow
I like the idea. Subaddressing with a + on my gmail addresses has been very useful, but I like the idea of it being a first class citizen.
Is there any chance that a hyphen will be rejected by some online forms? Even if it is is valid, there are some bad validators out there. I’m just thinking it might not be possible to use this for some of those services if all emails to your base address get bounced.
And bringing my own domain sounds interesting too.
I haven’t run into anything that rejects the hyphen. A few places reject the plus, like Microsoft. I’ve considered adding dot as a separator too. Right now, both hyphen and plus work, but that’s the only two. I have some future plans brewing in my head that might use the dot though.
Technically email to your bare address is accepted, it just isn’t usually visible to you. If you put “<bare>” into the search bar, you can see it.
This is useful for preventing general spam and whatnot, but it suffers from the flaw that if you sign up for my service (or more likely, sign up for a service I later get access to a data breach from) using an email like [email protected], I can instantly know that your bare address is [email protected].
A system like SimpleLogin/Proton Pass Aliases would be much better, where addresses are truly unique, don’t leak your original email in their names, but can be applied to any domain (including custom subdomains).
This could either be on the main domain, or to prevent cramming, a secondary domain just for aliases.
You could use both. Mine isn’t meant to keep your address anonymous, and that’s specifically because those kinds of services get blocked. But you can absolutely use one of those services that does forwarding with Port87, and I’d encourage you to do so. I wish I could provide anonymous addresses, but the fact that big companies definitely will block my domain if I do means it’s a no-go.
But you can also set up a “personal” custom domain with Port87. Then, every address on the domain becomes a label on your account. So something like [email protected] would be delivered to the “anon123” label in your account.
Also label names (what you see in your account) and label IDs (what creates the email address) can be different, so you can have a label named “Netflix” with an ID “or6neb5cij0f4hwocjf9idn”.
but the fact that big companies definitely will block my domain if I do means it’s a no-go.
They’ll do that with your regular domain regardless if enough people start using it, which is why I’m concerned this might result in your entire primary domain being flagged as a temporary email service.
For example, there’s no distinction from the perspective of the service to a domain creating aliases that are per-account, entirely random each time, and a domain creating aliases that use a standard format of yourname-service, because at the end of the day, users can still make unlimited emails for a given service (e.g. yourname-1, yourname-2, yourname-3, and they all sign up for the same site)
The reason why SimpleLogin, now owned by Proton, doesn’t offer proton.me addresses, and the reason why Proton, Tuta, and other email providers all limit the number of aliases you can make with the base domain to a set amount, and for paid users only (e.g. Proton limits you to 15 total, and you can only delete and replace 1 per year), is because, for example, if everyone could make unlimited emails on proton.me, then proton.me would get flagged as a temporary email service.
I’m not saying I hate the idea at all, I love to see more competition and useful tools coming to the email space when so much of it is dominated by just Google and a few more privacy-focused providers like Tuta and Proton, but I’m worried this mechanism will just get you flagged as a temporary email service by companies, just like most of SimpleLogin’s domains were on many services by default. (though they’re still quite usable on 99% of the web)
From the perspective of sites these emails will be used on, there’s no difference between how your domain acts, and how a temporary email service does, because no system exists that is implemented by all these companies to specially identify emails from your domain, know they’re using “labels”, and filter out duplicate registrations accordingly.
To do so would require scale, but you probably won’t get scale unless you can avoid getting flagged as a temporary email service, which won’t happen unless those services all had such a system in place to the first place.
I see what you mean, but the same would apply to any service that allows subaddresses, like Gmail for example. Yourname+1, yourname+2, yourname+3 on Gmail all go to your address. Many email services allow subaddressing, and some even use the dash, like Port87. In terms of the addressing format, there’s nothing new about Port87. It’s everything that happens behind the addresses that’s new.
However gmail is a large, incredibly well known service, and many sites understand that the + on gmail specifically is for subaddresses and will deny using the same email with subaddresses different times.
Contrast that with just using dashes like Port87, and most systems don’t have anything made to parse for dashes, as it could then result in problems where an email service like Gmail, or any other provider out there, allows people to put dashes in their base email, and someone can effectively block someone else from signing up for a service by making a new account named theirname-1, signing up, then the service would think that theirname-1 is also owned by theirname, and block theirname from signing up later.
The + is a relatively well known standard for email subaddressing, but dashes are primarily used by people just inside their email addresses instead of a space, for example. Thus, most server side implementations will never be configured to understand dashes as indicating a label, specifically for your domain, they’ll just see a large volume of constantly created new emails, that act like a temporary email service, and assume you’re one.
This has the same problem as before, where you’re not large enough to justify being specially considered by login pages that will understand what your labels are, but are also not going to get to that scale if you get filtered out as a temporary email service.
I’m probably going to stop responding now, as I think that’s about all I can contribute, but I’d just say that if this is the exact mechanism by which you plan to implement subaddressing, make sure you’re frequently checking any widely used blocklists online for temporary email domains, because someone will probably end up submitting your domain there at some point based on the behavior of the service, and it’s incredibly hard to get off once you’re on. (and consider making a page on the site explaining why you’re not a temporary email service, like SimpleLogin has)
That’s a good idea. I’ll add that to the website to explain why it shouldn’t be used as an anonymous or temporary email.
Yeah but emails to the bare email gets rejected as far as I understood
Even if they do, most people will use identifiable service names, so then you’d just have to brute force with yourname-google, yourname-facebook, yourname-lemmy, yourname-bank, etc. Still a major risk compared to traditional aliasing services in my opinion.
I like this, and I’m willing to give it a shot.
- Is the app available on F-Droid?
- Can I count on this to be a long-term (10 yr+) solution to move away from protonmail for good?
- Where are the servers geographically?
- Can I add and manage filtering rules?
- Can I access port87 mails with a standard IMAP client like k9 or Thunderbird?
- Do you allow users to send mail from SMTP sources like, for instance, sendmail or postfix on my server so I can send myself alerts?
Edit: I’d be happy to pay for mail as a service if it works for me.
Right now there’s only the progressive web app, but I will be working on a native app shortly, and it’ll be on F-Droid. It definitely will not rely on any Google closed source stuff. I use /e/OS on one of my phones, so I’ll be testing it on that.
It’ll be around at least as long as I am, and my wife and dad both have instructions on how to pass it to my chosen successor if I pass. If you want to make sure not to lose your addresses though, I’d highly recommend using a custom domain. That way if something happens to it, you can move to another provider and keep your addresses.
The servers are in my office in Temecula, CA, USA. I’d love to host it internationally, but that’s not something I can work on right now.
It depends on the filtering rules you mean. Each label has settings for notifications, mark as read, and show in the Aggbox (like an inbox, in the more traditional email model). They can also be blocked, and have screening. Beyond that, there’s no additional filtering, but that is planned. I don’t know if I can technically support full sieve scripting, but that’s my goal.
Right now, there is no IMAP/SMTP support. The code for SMTP support is done, but not IMAP yet, and I’d like to launch them together. I’ll be working on IMAP very soon (right after mailing lists and aliases). You can install the web app as a PWA on your phone, though, so it’s a great experience on mobile.
Feel free to try it out and let me know what you want me to prioritize. It’s my full time job, so I’ll be actively working on improvements.
Firstly—this project sounds cool! I host my own mail server already so have no need for your project, but it’s good that there are options out there for people.
The servers are in my office in Temecula, CA, USA.
In the long run I would really suggest you get servers in another jurisdiction—USA will be a no-no for a lot of potential customers.
Good luck though :)
Believe me, I know. I don’t want to be here even. If I get enough income going that I can pay for hosting costs, I’ll split the server into US and somewhere else (maybe Europe), so you can choose where your data stays.
You can rent offshore VPSes for pocket change. There are some hosts that don’t block mail ports.
There’s a bit more to it when it’s email, but it’s just not something I can prioritize right now. My costs are pretty low because I host everything on my own hardware, and adding rented hardware will have to come later when it makes more sense.
It definitely will not rely on any Google closed source stuff. I use /e/OS on one of my phones, so I’ll be testing it on that.
eOS is using microG
Ah, ok. I’ll make sure to test it on a bare AOSP build then. Thanks for the heads up. :)
Wow, what a great response!
OK, you’ve sold me, I’ll sign up and try it.
Your personal willingness to explain your philosophy of the service is what swayed me, and I think that will go a long way in other forums and sites to pick up new subscribers.
Thank you. :) Let me know what you think and how I could improve it.
I like the idea! Keep at it!
Looks very interesting, and the pricing is pretty decent
soo… what’s stopping spammer mcspammer to send emails to [email protected] and thusly hitting your eyeballs? or spam you with millions of new tags by adding random junk to the hyphenated part?
The screening on the friends label would stop them from emailing that one. Any other labels would just become pending labels. If you’re not expecting a new label, there’s little reason to check your pending labels anyway, so you probably wouldn’t notice them. They get deleted after 60 days if you haven’t approved them.
There’s nothing stopping an actual human being spammer with a real email address from reaching you, but that’s kind of the point. Real people with real addresses can reach you. Automated spam has a much harder time reaching you. They’d have to have a custom system designed to target Port87. Even then, I could just add something like Cloudflare Turnstile to the screening process.
Spammers don’t typically like to use real email addresses with a real return path because they get shut down quickly.
Has it got a calendar by any chance, or is one planned?
Dumb question Why is calendar or contacts expected to be bundled with email service? Is there any technical limitation or advantage doing so?
Contacts makes sense for email address autocompletion, but there’s no need for that to be bundled with the email service itself given that email address completion happens on the client not the mail server. And calendar I guess people can send emails with ics links, and most mail clients let you accept or decline invitations from within the client, but that connection is more tenuous. Maybe it’s just me but I rarely receive event invitations over email, and I never use the accept/decline feature; I just add it to my calendar manually.
For contacts and calendar syncing I just use Nextcloud with CalDAV/CardDAV clients on all my devices.
I wouldn’t say it’s a dumb question. Calendars being built into an email service makes invitations easier. Right now, if I want to invite you to an event, I would create an event with my calendar service, then download an ICS file for the event, and attach it to an email to you. If the calendar is built in, the invitations can be automatic.
Also, with Port87, I could make calendars have their own addresses. So, I could have all my work stuff added to one calendar, and all my friends stuff added to another, based on the address the invitation was sent to. I think there are solutions like this already, so that wouldn’t be a new concept, but it would integrate well into Port87.
No calendar yet, but I definitely plan to make one in the future. My current todo list, in order, is:
- Aliases and mailing lists for custom domains.
- IMAP/SMTP support for third party email clients.
- CardDAV support for contact syncing.
- Native mobile app.
- Calendar.