I recently launched a new email service called Port87. It's very different from how current email providers work. It's all based on tagged addresses, aka subaddressing.
My email is hperrin@port87.com.
When I want to give my email to a service, like Netflix, I wouldn't use that though. I'd use something like hperrin-netflix@port87.com. Then when Netflix sends an email there, Port87 will automatically create a pending "netflix" label for me to approve. All of Netflix's emails will go there. Each email address is separated from all other accounts, the email is automatically organized, and you can control whether you want notifications and to mark new email as read per label.
Then if I want to give my email to a friend, I would use something like hperrin-friends@port87.com. My "friends" label has screening enabled, so any new sender gets an email back asking them to click a link to prove they're human before their email is delivered.
Finally, I can share hperrin@port87.com publicly, because any email sent there gets an autoresponse with a list of addresses for my "public labels". The sender can choose which label is best for their message. For example, I have a lot of open source projects, so I have a public "opensource" label.
I've been using it for over a year now since I launched it publicly, and it's really a night and day difference over the more traditional style email services.
Personally I'm using Migadu at the moment and for me the killer feature is that the price per month is not bound to users and/or domains. I have several tens of "users" (many are myself but for reasons with completely separate identity as far as the server is concerned), and 8 or so domains. There is barely any traffic (in the order of <10 emails in and out per day).
Most other providers I would have to pay 10-30x as much, just because I have several users and domains, at a grand total overhead cost of 0$ to the provider.
I'm going to give you a niche idea. An email provider / hosting / client focused on a single user owning a whole domain.
I don't care about individual email addresses. What I really want is to be able to handle asterisk@asterisk.example.com both for receiving but more importantly for sending.
Receiving is fairly easy with aliases and catch-alls. But for some reason all clients / providers make it really hard to send emails from any address you own.
I think there is money in this problem.
For me, privacy is the most important aspect.
- don't train AI on my mails
- don't sell my data to advertisers.
- protect my data from overzealous law enforcement
Right now, I'm running my own mail server on a dedicated machine. I wouldn't mind getting rid of it. One feature i need is a configurable sub-adressing separator (not just "+").
I think so. Another recent entry is purelymail, which I love because it's a flat rate for unlimited domains. I'm trying to spin up multiple "small bets", and per-user pricing is unreasonable, and for better or worse, I don't want a single catch-all email.
A secondary benefit for me is that Purelymail is cheap ($10/yr). But a major downside is that the CalDAV implementation doesn't yet support Accepted/Rejected scheduling responses. It's fantastic that they have CalDAV at all, but this one feature would seal the deal for me.
I don't mention it to suggest you could directly compete on that unlimited-users value prop, but it's an example along the lines that max_ suggests.
One that comes to mind is integration with recent ActivityPub / fediverse stuff. There might be cool design space there
I think email today "just works," esp. since spam filtering became decent and auto-categorization. I don't think being a new email host is where the win could be but at the client end. That's what recent successes like Superhuman, Hey have been.
Until Thunderbird added support for Exchange there was a slight hole on Linux (no FOSS mail solution with support for it) but right now I'd say everything works more or less out of the box. SPAM filtering and categorization could be a bit more automated, but in general I'd say everything is covered.
Mail server projects are a different story though. There's no simple way to create a fully fledged mail server without weeks of pain to configure spam, routing, calendars, security... there are some preconfigured images in all major cloud platforms, but nothing like `apt install mail-server` that pops out a simple config assistant in an already existing server.
Does anybody remember Eudora? Not the Thunderbird version, the one before that. It was fast as hell with cool shortcuts (eg Alt-clicking on from groups all the emails from that sender) and therefore super easy to find stuff and organize your email even if you had way too much of it. My M1 MBP is probably 2000x faster than last computer I had running Eudora, but modern email is still slower and clunkier. A web based email service that is blazing fast and makes sorting/filtering/searching my email easy is something I would be happy to pay for.
Another email service I want would be one that I have to authorize the sender before their mail gets into my inbox even if it's not classified as spam. Sure, that email can be routed to some dumping ground so I can sift through it if an expected message is missing, but by default nothing goes to my inbox unless I've routed the sender there. And when I "accept" a sender it should be trivially easy to route them to any mailbox or make a new one. And it should be trivial to manage those routing rules later. All this is already possible with existing services, but enough of a pain that I don't maintain it.
The last thing is a tough one: a way to port my existing email into the system and a guarantee that I won't lose my email if you shut down in a few years. Trying other email providers is not worth it if I won't be able to find anything from the past and I might be at risk of losing them in the future.
An email provider with a high quality, efficient native app that has first-class offline support, threading, some well-thought-through AI-based categorization of emails, respects my privacy, and does not issue network requests based on email content would be quite welcome, IMO.
Look at Fastmail for a surprisingly popular service that gets most of this massively wrong.
I would not exactly expect innovation.
I would expect: fair pricing, local (as in the country you live in) hosting (for legal reasons), features like using your own domain, maybe service for calDAV cardDAV etc. to tie up the bundle.
Also no user data aggregation, no excessive logging, etc.
Really, the one thing I would look for to recommend this to others is having that _and_ ease of use for non-techies.
I think Inbox was loved by many until Google shut it down: https://en.m.wikipedia.org/wiki/Inbox_by_Gmail
I've also seen the idea of an email client organized as Facebook - with feed, groups, comments, DMs etc, but where all the data is sent as emails
I could be wrong but I think a good way to have an edge is pricing. Email is generally a low bandwidth service with lots of cold storage, yet for most personal email services the price doesn't seem to reflect this.
Like for example Google One is $2/100GB and $10 for 2TB yet you can definitely get storage much cheaper than this, and bandwidth is close to, if not free on bare metal clouds.
I've been considering this myself. That or some IRCv3 offering.
If you go the commercial route you'll almost certainly end up building an ERP/CRM, as it's what a lot of companies seem want: a personalized ERP/CRM (not sure why, though it does kind of feels like just outsourcing an IT dept whenever this has come up in the past)
My wish list: 1. iOS native sync, preferable without Active Sync. This includes syncing of read states. 2. Custom domains for families and small business without shelling out $10 per user and month. 3. A shared mailbox within the custom domain as „family“ contact mailbox. At least two people should be able to read it, possibly sync read states AND answer from that address, including sent mails need to be stored in this mailbox. 4. A very simple way to setup the accounts on different systems, especially iOS.
AFAIK, except 4 those are already solved, but not within one provider - without paying those mentioned $10/u/m.
There's room for one with better privacy and, hopefully, E2EE -- if you can find a way to make it work.
There was a thread about this just a couple of weeks ago: https://news.ycombinator.com/item?id=40924055
Protonmail is increasingly Gmail-like in every respect -- with its AI "tools," with it asking you for personal information upon account setup, and in obliging cooperation with foreign courts -- so a new service which doesn't deviate from a focus on true privacy would rapidly find its niche.
I want a "Mail-in-a-box"/Plex mashup... that lets self hosters setup a domain, with a mail server than handles all the spam and "trust" shit that comes with hosting a mail server.
The new hosting service does all the trust and anti spam/virus/phishing stuff, and maybe does or helps with the neat stuff you get from gmail (indexing and clever integration to calendar and travel/shipping integrations.) We'll either buy VPS or self host on hardware in our house the actual data.
I replaced Outlook with Obsidian + syncthing. It would be nice to add email into mix. Some sort of service that fetches mails from multiple sources, and exposes them as filesystem. I want to link individial emails from notes.
Am I the only person who thinks email search is completely useless? I’d happily pay for a service that lets me search selected fields for a substring (eg title only, title and body). Give me regexes and I’ll pay more.
I need server-side filtering rules that I can configure with a web UI. That basically leaves only Exchange and GMail. AFAIK, there's no privacy-respecting small provider offering server-side filtering rules.
Email is root node for almost all services. Some future improvement may be a root node and also email -- but not necessarily bundled as a unifed item. Maybe it will have some web3 stuff in it or similiar.
I would say yea, if it is going to be a replacement for the Google Workspace. I hate it.
For personal use I feel like we already have a very good set of email providers. For small businesses not really.
Interested in having a private conversation about this? Have a couple of ideas I’d like to share if you want to leave your contact details.
I think there could be, sort of. I am currently using Migadu, which works great and I don't plan on leaving. Their selling point is that the amount of domains you add is not limited. It's affordable and reasonable, and a great option as indie dev with multiple projects. No fictional per-seat pricing that does not correlate with costs to the provider.
What I think is outside their target market, but what would be a great next step for projects with slightly more needs would be a Google Workspace/Outlook sort of service with a similar approach. Simple CRM features, shared contact management, calendars, signatures, etc. Roughly what Proton is doing without the Silicon Valley pricing, or maybe even Hubspot.
Bring back the original GMail experience.
[dead]
[dead]
There is always room for a new product if you provide value that established entities don't offer.
Proton Mail for example built it's business around privacy.
There is alot email is capable of that has been forgotten.
1. For example I am part of this mailing list that is like 30 years old. And it feels sort of like a small niche social network. There is alot of improvement that needs to be done in this space.
2. Alias generation. I really would like to subscribe to some news letters or place my email out in public but I don't want to do it with my real email for fear of being spammed.
3. Cataloguing & curating emails. It would be really nice if I could book mark different messages into different folders.
4. Read receipts. All email hosts claim to have this feature but I have actually failed to use them for every single service I am signed up to.
5. Make it easy to configure custom domain names.
6. Tagging doesn't work on Gmail's mobile app. It's so badly implemented on other services.
7. Have a good API why do I have to go to sendgrid?