Wouldn’t trust it at this point … but what I would trust is a blank router that has a GitHub repo with a containerized build that I could run on my own machine and something ai ish that lets me flash it :)
Would be legit to be able to have an Llm do a source audit before installing …. Not perfect but damn sure better than nothing
My idea is one step further: selling the device with various ad-blocking and user-selectable tiers of security configuration included. Stop your devices spying on you, detect/notify if you have any infected devices on your network. That kind of thing.
Similarly though, there's limited appeal to those who aren't already capable of setting it up themselves.
I'm still slowly pursuing this because I believe in it, whether it's profitable or not.
ISPs would have been able to do this kind of thing since forever but, strangely, I don't see it advertised as a potential value-add at all. Had ISP's commonly taken up such a mantle, players like NextDNS probably would be less necessary.
I used to use dd-wrt, but I later settled on ASUS routers, and installing "ASUS Merlin" firmware on them.
I find ASUS hardware high quality in general. And using an open-source project like ASUS Merlin means the pain-points and limitations that I would face with stock firmware have been resolved to (at least to satisfactory extent) when I need reach for the hacker type configurations.
There are some Chinese companies making more open router dev boards like Banana Pi R4. You can install openwrt out of the box. But generally, the SoCs that power the current generation of routers are locked down / highly proprietary (and always have been). It's the chips, not the boards, which matter most.
I don't think it will work as a business. The people who want dd-wrt will just install it themselves, not pay additional for it to already be installed. However, I do think you may be able to sell an ad-blocking router with ad-blocking already installed. I think people may consider that enough of a feature to pay for.
Agree - the firmware is never really a selling point.
But a shop built around routers could have a few angles:
1) Privacy/ ad-protection native (as you say)
2) High security and performance (fully tuned DD-WRT). Even solving bufferbloat for end users would be amazing - the world is barely catching up on that.
3) Good software add-ons that aren't proprietary junk. Asset management/upgrade, sign-in/landing pages, monitoring and trafffic shaping that holds up, etc.
Bunlding FOSS to solve these problems well, even on some else's hardware, does seem like work, and work can be charged for.
I'd be happy to be proven wrong, but I think that routers are a pretty geeky thing that most people don't want to hear about. I had offered to buy and set up a pfsense router for my parents for a few years ago and they were decidedly not interested, even after I tried explaining the merits of it. Pretty much anyone who is going to be interested in DD-WRT is going to be someone who can set it up themselves.
pfsense can sell for a relatively high price, I think, because they market themselves more towards businesses and big corporations.
Interesting related question -- I recently inherited several WiFI AP/Routers that are openwrt or DDWRT compatible. Are there any packages that folks are working on, or know of that can centrally manage a fleet of them?
If there wasn't, I was going to use this as a project to explore vibecoding...
Alternately, what's the problem you're trying to solve? Who are your potential customers? Why do existing options in the market not address their needs?
As the other commenter said - but if you loved IOS before it was iOS you’ll find routeros comfortable if different, if you like rc.d you’ll like DD-WRT.
they're entirely different beasts. dd-wrt is more or less a normal linux distro that targets embedded devices. it has a "unified config" that abstracts away configs for many packages, but those abstractions are leaky. when you log in by ssh, you're using a shell and editor to edit config files.
compared with RouterOS where the Linux bits are almost completely hidden behind the proprietary config engine (and the proprietary cli and webuis that power that config engine). for the most part with routeros, you'd be hard pressed to know that you're even on a unix let alone linux.
I used to use DD-WRT, and Tomato, and Shibby's mod of Tomato, and Advanced Tomato, and other stuff. Somewhere, I'm sure I still have a WRT54GS that I hacked ~20 years ago to accept SD cards. It was a fun time in using inexpensive hardware in unintended ways.
At one point, and I don't remember exactly when, I just started using OpenWRT instead. It wasn't always the case, but it presents a very usable GUI straight away these days and it has a ton of flexibility for getting creative later.
At home I run it on a Pi4 that I bought for the purpose. It keeps up with my 500/20Mbps DOCSIS internet very well indeed. I've used it with a USB NIC to get dedicated ports, and also with VLANs and a managed switch instead. I just download and flash the card, plug stuff in, and configure it. No real issues -- for years, now. (Wireless is provided by Mikrotik access points.)
When something faster happens in my neighborhood (which may be very soon; I had a very brief chat with a dude whose fiber splicing trailer was parked across the street just yesterday), I'll probably want to move to something more capable.
That could be in the form of faster Pi, or an old PC running OPNsense -- or even stick with OpenWRT. No big deal.
If I really wanted a tidier package for light-weight travel or something, then I guess I'd pick something appropriate from glinet and put an official OpenWRT image on it.
For heavier-travel, I've got a kit with a Pi4, a managed PoE switch, and a backup Mikrotik box (that can be a switch or a router; it runs OpenWRT as well and the normal config is that it just provides more switch ports), and all this stuff is mounted onto a 1U rack shelf that is in-turn inside of a road case. It's meant to be physically durable and to include enough gear that if any part of it dies then I can replace that functionality Right Now without spending time finding parts in the middle of nowhere.
So, no: I would not buy one that was prefabulated. It wouldn't scratch my itch.
---
For everyone else? Honestly, the regular folks I know who ask me questions about networking aren't into that stuff at all or they wouldn't be asking. They don't care about ultimate speeds, or latency under load, or any of those things. They don't care about running Internet services from their network, or services on their local network, and probably don't even have a printer to abuse if something goes awry. They're also not interested in exploring the bowels of their home to install cabling for PoE access points.
They just want decent-enough wifi coverage that they don't have to think about it, and then: They want to succeed at not thinking about it.
So for those folks, I've come to recommend proprietary multi-node mesh systems. It doesn't really matter to me which one it is, or how terrible some aspect or other is: They all tend to check the right boxes, they update themselves, and then: Nobody thinks about them.
IMHO, those who want a router with open-source firmware are going to have the technical knowledge to do the finding and flashing themselves, by virtue of being embedded in tech deeply enough to recognize the value of an open-source firmware in the first place.
Those who look at you blankly when you mention open-source router firmware aren’t likely going to become customers.
You would be surprised at how many people that know enough to be curious but are not ready to jump straight into flashing custom firmware yet. I've seen people sell pre-flashed wired and wireless routers like this for years on platforms like EBay. I think there's already a niche there, even if it's small.
I'm not sure how approachable tools like Claude Code have made this for people though. If I was just getting started, I would probably just have the tool do it for me.
Would be legit to be able to have an Llm do a source audit before installing …. Not perfect but damn sure better than nothing
Similarly though, there's limited appeal to those who aren't already capable of setting it up themselves.
I'm still slowly pursuing this because I believe in it, whether it's profitable or not.
Edited to add: Part of what I'm setting up: https://lists.uninvitedactivity.com/DNS/, a vague explainer here: https://news.ycombinator.com/item?id=49990394
ISPs would have been able to do this kind of thing since forever but, strangely, I don't see it advertised as a potential value-add at all. Had ISP's commonly taken up such a mantle, players like NextDNS probably would be less necessary.
I find ASUS hardware high quality in general. And using an open-source project like ASUS Merlin means the pain-points and limitations that I would face with stock firmware have been resolved to (at least to satisfactory extent) when I need reach for the hacker type configurations.
https://www.gl-inet.com/en-de/products/gl-be14000
so.. yes? I haven't gotten around to flashing stock openwrt on it - but I probably will at some point - seems to be supported now?
https://forum.gl-inet.com/t/flint-4-openwrt-vanilla/71383/10
I think it would be difficult to compete with existing offerings - even though they are quite few.
1) Privacy/ ad-protection native (as you say)
2) High security and performance (fully tuned DD-WRT). Even solving bufferbloat for end users would be amazing - the world is barely catching up on that.
3) Good software add-ons that aren't proprietary junk. Asset management/upgrade, sign-in/landing pages, monitoring and trafffic shaping that holds up, etc.
Bunlding FOSS to solve these problems well, even on some else's hardware, does seem like work, and work can be charged for.
pfsense can sell for a relatively high price, I think, because they market themselves more towards businesses and big corporations.
No, I'd rather flash it myself and use OpenWrt instead of DD-WRT.
If there wasn't, I was going to use this as a project to explore vibecoding...
Buy https://www.gl-inet.com/en-us/products/gl-mt3600be and move on with your life.
Alternately, what's the problem you're trying to solve? Who are your potential customers? Why do existing options in the market not address their needs?
compared with RouterOS where the Linux bits are almost completely hidden behind the proprietary config engine (and the proprietary cli and webuis that power that config engine). for the most part with routeros, you'd be hard pressed to know that you're even on a unix let alone linux.
At one point, and I don't remember exactly when, I just started using OpenWRT instead. It wasn't always the case, but it presents a very usable GUI straight away these days and it has a ton of flexibility for getting creative later.
At home I run it on a Pi4 that I bought for the purpose. It keeps up with my 500/20Mbps DOCSIS internet very well indeed. I've used it with a USB NIC to get dedicated ports, and also with VLANs and a managed switch instead. I just download and flash the card, plug stuff in, and configure it. No real issues -- for years, now. (Wireless is provided by Mikrotik access points.)
When something faster happens in my neighborhood (which may be very soon; I had a very brief chat with a dude whose fiber splicing trailer was parked across the street just yesterday), I'll probably want to move to something more capable.
That could be in the form of faster Pi, or an old PC running OPNsense -- or even stick with OpenWRT. No big deal.
If I really wanted a tidier package for light-weight travel or something, then I guess I'd pick something appropriate from glinet and put an official OpenWRT image on it.
For heavier-travel, I've got a kit with a Pi4, a managed PoE switch, and a backup Mikrotik box (that can be a switch or a router; it runs OpenWRT as well and the normal config is that it just provides more switch ports), and all this stuff is mounted onto a 1U rack shelf that is in-turn inside of a road case. It's meant to be physically durable and to include enough gear that if any part of it dies then I can replace that functionality Right Now without spending time finding parts in the middle of nowhere.
So, no: I would not buy one that was prefabulated. It wouldn't scratch my itch.
---
For everyone else? Honestly, the regular folks I know who ask me questions about networking aren't into that stuff at all or they wouldn't be asking. They don't care about ultimate speeds, or latency under load, or any of those things. They don't care about running Internet services from their network, or services on their local network, and probably don't even have a printer to abuse if something goes awry. They're also not interested in exploring the bowels of their home to install cabling for PoE access points.
They just want decent-enough wifi coverage that they don't have to think about it, and then: They want to succeed at not thinking about it.
So for those folks, I've come to recommend proprietary multi-node mesh systems. It doesn't really matter to me which one it is, or how terrible some aspect or other is: They all tend to check the right boxes, they update themselves, and then: Nobody thinks about them.
Those who look at you blankly when you mention open-source router firmware aren’t likely going to become customers.
I'm not sure how approachable tools like Claude Code have made this for people though. If I was just getting started, I would probably just have the tool do it for me.