I prefer to assume he's lugging a 12U mini relay rack around with him, including a UPS and cellular fallback for networking, so he can run an Android VM and use it as his cell phone.
To me, "my server is now..." implies there was some server, it used to do server things, but now it has been repurposed to do something else.
When I read the title, I literally thought "wow, I've seen a lot of people make servers out of old unused phones, but this is the first time I've heard someone make a phone out of an old unused server!"
Some newer phones support bypass charging, which powers the phone directly instead of continuously cycling the battery. Mine doesn't support that, unfortunately, so I've limited charging to 80% to reduce battery stress.
That's surprising to me. It's been a standard feature on laptops for 20 years.
Edit: Or I should say, it was 20 years ago. Maybe modern stuff is worse, but my M2 MacBook Air and Framework 13 Intel Core both power on without a battery.
LiPo batteries will fail over time if left fully charged. If you don't need the battery it's best to keep it somewhere around 40%. If 80% is the only option, that's usually good enough.
The primary failure mode starts when the cells are charged near their maximum voltage and kept there.
Afaik mobile electronics’ power circuitry is designed so that power goes straight from charging port to the processor, without constantly charging/discharging the battery, once charging is complete. Unless the system suddenly draws a lot of power (more than the charger is supplying) - which is unlikely in a smartphone.
This is true but lithium-ion batteries degrade even while not being charged or discharged, and this is worse at the extremes of the state of charge (including fully charged), and at higher temperature (and often a plugged in device is on constantly and therefore spends more time at a high temperature than it normally would), so it's still not great for the battery. A lot of devices have a mode where the batter is kept at a lower state of charge if it's been plugged in for a while to try to reduce this.
Some phones bypass the battery when plugged in, but not all do. I know it's advertised as a feature of some gaming phones because graphics-intensive gaming is one use case where you'd want to use the phone heavily while plugged in.
Even then, some phones that advertise bypass charging are actually using a software-based implementation, and they can still end up with swollen batteries after being plugged in for extended periods.
Steamdeck lets you set your preferred % and the dock does pass through charging/power, which I love because that combo guarantees the longest possible life for the battery. I turn it back to 100% for flights, otherwise leave it as 80% at all times. Mine is almost 4 years old now but I don’t feel even a hint of battery degredation yet.
> It will probably destroy this battery very quickly.
'Very quickly' seems to be measured in decades from what I see
Most people I know have their laptop docked, either near-permanently at home or taking it between home and a desk with dock (so it's nearly always plugged in when in use)
The 'servers' I've had were all old laptops - yes, with battery (don't have to buy a separate UPS if one is already included). I've now also had two non-server laptops that were plugged into power pretty much all the time. I treated my first laptop perfectly: taking out the battery anytime I'd be in a classroom for several hours and it was already full enough (you know, back when sliding out your battery was possible in the first place), but I'm not noticing any faster degradation on the systems I treat poorly. I'll activate the 80% cap function where it exists (that is, in the latest phones and my work laptop's UEFI config panel), but the evidence of it being helpful has yet to materialize
I have read that studies show that charging to 100% causes degradation and I trust those more than random people's anecdotes (and I'm a random person), but I'm starting to wonder if I've maybe just misunderstood the study results. Maybe it's like how smoking increases your chance for some cancers but you might also be fine in the end. I see people with plenty of much younger and better-treated batteries that can't sustain literally 30 seconds of use whereas mine are reasonably close to the new state (some haven't had many charge cycles, just sat at a stressy high voltage all the time, which was indeed supposed to destroy them), and I don't know how to explain that otherwise
What you can do, with some extra steps, is have Home Assistant installed (and have a Home Assistant server, I guess!) on the phone and use a smart power switch that's also associated with Home Assistant.
The phone's battery can be reported to Home Assistant, where an automation can be triggered to turn off power to the phone when the battery gets above a certain percentage and turned back on when it goes below.
Not the best solution, but certainly an alternative from messing around with trying to get the phone to work without a battery.
Most phones won't boot without the battery. Obvious workaround is to connect a power supply (and large cap) directly to the BMS, but in my experience most phones will shut themselves down when they think they've drawn more than the capacity of their battery, even if the voltage is still high. You can sometimes get around that by rebooting the device every couple of hours (and there might be a software/firmware fix in some cases?), but it's a lot less annoying to just accept a small risk of a house fire.
You could throw the phone in an ammo canister, if the wifi AP is close enough. When I was building hobby quadcopters, which requires several ~half-pound batteries that are a chemistry optimized for maximum power density, a lot of guys recommended steel ammo cans for battery storage.
Depends on how well the canister insulates it. Steel conducts heat fairly well; it'd get hot inside but I'd guess some airflow around the canister would probably keep it from getting hot enough to cook the electronics. (definitely don't take my word for it lol)
Phones are very low power, likely just a slightly warm spot on the can that radiates, conducts, and convects the heat out. You could also use one of those peltier module phone coolers, to ensure the heat gets out of the phone and into the air faster.
> but it's a lot less annoying to just accept a small risk of a house fire.
Agree! But do get one or two of those "small automatic fire extinguisher" or whatever they are, that you can attach on stuff, for electrical panels and what not, and put right next to it together with a traditional fire alarm, and you're good to go :)
FYI you can't do this on phones with locked bootloaders as well as the author did. you could use termux but if termux runs on anything lower than android 8 termux will be significantly limited and you will need postmarketos and maybe fix support for it. with termux, at best things will be slow without root, as the author mentioned rooting resulted in a speed increase.
without root, you also can't bind to ports on any android phone, and rooting often requires an unlocked bootloader and not all android phones support bootloader unlocking. You also need an unlocked bootloader to install postmarket os.
so without binding to ports you can't run a web server unless you run it through a proxy server which the author did, or containers at native speed on postmarket os because you will not be able to flash the phone with postmarket os if you have missing kernel syscalls on android nor root the phone for the speed increase on android depending on your android kernel version i assume.
so with android containers may or may not work without emulating x86, and without root chances are it will be slow specially if you are forced to emulate it as x86 because it turns out you have missing syscalls, nor root to make them faster if they exist on android. again no chances to install postmarket os nor rooting to make containers faster, without bootloader unlocking.
I still think the best bang for your buck is an old desktop PC. Something like an i5 with 8GB RAM and an SSD will suit 99% of home server uses with capacity to spare.
Even in this environment of inflated hardware prices, you can get one for a reasonable price.
Although I do see the appeal in getting unconventional hardware to do what you want, just for the love of the game.
It only makes sense if we are measuring raw capability per dollar. Once you factor in noise, physical size, idle power, battery etc, I think the calculation changes.
For a home sever that's mostly idle, it is almost like commuting in a pickup truck.
It averaged 69w. The Mac Mini I replaced it with uses just under 5w.
A Mini is possible the worst bang for buck. Also, the horror that that is Desktop Docker can’t be over emphasised. I know there are better ways of using Docker on a Mac, but having got something working I can’t face rebuilding it.
My server is an extremely quiet Xeon workstation: unlike 1U servers, workstation are typically quiet. The good thing with a Xeon workstation is that it supports ECC RAM only and there are plenty used. For there's no market for used Xeon systems, so you can find stuff like 22-cores / 44-threads CPUs for $25 or something.
Some mod Xeon systems to use regular PC PSU like Be Quiet! ones and Noctua fans, etc. but my HP Z440 workstation is so quiet already I didn't need to.
In my case it's an headless server and is sits between the vacuum cleaner and the cat's food: there are literally only two cables connected to it, the power cable and the ethernet cable.
I can live with the footprint of one tower.
I run many VMs and containers inside those VMs: Git server, pastebin, NAS, containerized/VM-isolated AI agents, etc. but I don't need it to be up 24/7. I basically only need those when I'm working. So electricity usage ain't a big concern to me as it's more like running 8/7.
I don't feel like I'm commuting in a pickup truck: I feel like I have the perfect machine for the task.
I just bought a refurbished Lenovo k14 with 32gb of ram and Corei7-1165G7 with 500gb for 350$ which I'll use as my 2nd home server.
My first home server is a 2nd hand gaming PC that I bought for 500$ with 32gb and 7700rx video card that I use for GPU based tasks but it's quite audible when it's working so wife and I agreed it has no place in the living room anymore and is being deported to the corner of my office.
You can still find really good bargains for home servers from refurbished good computers, and given how well solved tech reliability is now, it's really not worth buying new.
This is my situation exactly. Got an i5-4570 with 16GB, no GPU, running Unraid with a bunch of containers for Jellyfin, qBittorrent, Crafty Controller (minecraft), HomeAssistant, etc.
I'm using a PCIe expansion card to get double NVMe drives for the cache disk, and then three (up to five) 3.5" spinning 8TB drives for long term.
I tried really hard to get it on 32GB but sadly the motherboard only has two slots and only supports 8GB DDR3 modules. I literally ordered some of the rare late-generation 16GB ones and it refused to boot, with no firmware updates available from the vendor (Advantech, board is an AIMB-274).
> I tried really hard to get it on 32GB but sadly the motherboard only has two slots and only supports 8GB DDR3 modules.
Buy a refurbished Xeon workstation with DDR4 RAM. If you're in the US or China they're basically free and if you're in the EU, they aren't too pricey (like you can definitely find ones for less than 200 EUR). Slap your double NVMe expansion card, your two NVMEs and your 3 3.5" spinning rust in it. If the workstation doesn't come with a Xeon with lots of cores, just buy one on eBay for $25.
I’ve wondered for a while if they could be used as an extension to your laptop like the old 80s PCs where you would plug in hardware.
So ‘make it a server’ is the obvious way I guess, and then you could do things like render webpages on there and then stream them as images to your actual computer (like the now shutdown Mighty app), or run ffmpeg on there to convert a video, etc
Cool project, but slightly better: before the rampocalypse I bought one of those aliexpress fanless linux boxes designed to be a router. I put a 32GB stick of ram in it and a 250GB SSD. I run opnsense in a virtual machine to serve as my home router and then I put the rest of my home server stuff on other virtual machines on that same box. It's still fanless/silent. It takes up zero additional space because it is my router. It would already be running 24/7 because it is my router so the additional energy usage of an additional VM is likely minuscule. The latency to my router is basically zero since it is on the same physical box. And it is using a real SSD so it has faster storage with almost certainly better endurance.
Those saying it's said backwards clearly aren't self-hosters!
iPhone hardware wise is like an amazing alternative to rpi (way better hardware)
But the software makes no sense at all for a server, with screen locking due to inactivity and whatnot, not to mention energy and network priority is designed around a consumer use case on the move.
If you run Linux on it and take advantage from the Apple hardware like the light sensors and accelerometer - maybe it makes sense for measuring wave height, sky color, attached to a buoy or oil rig or something, you can even show it on a map due to its GPS. Assuming the oil rig has WiFi or a cell phone booster.
Idk though, maybe for something that important you just get the right stuff - cool hobby nonetheless.
The main argument you give for it not making sense can easily be changed in Settings. The energy and network considerations have no impact on running a virtual machine in UTM – iPhones will give everything they can as long as thermals are fine.
Can you please tell me more? I'm trying to build a self hosted application (with Claude) so people can scan a QR code and place orders / send messages to the kitchen without having to flag down a server / waiter. It is supposed to save time and money for the restaurant and for the patrons and customers it is supposed to be more convenient. Do you not like this option?
Every system like that I've ever used has three fundamental problems:
1) It's optimized for the restaurant first, and the patrons second or not at all, and it shows.
2) Someone thinks "hey, we could sell ads on this", and now it's even more a piece of garbage. (This is also what happened with all the little kiosk-tablets on the tables at some restaurants. We just set them face-down on their screen and talk to a human.)
3) Most of them never anticipate unexpected problems. They might have an "order X" button, but they often don't even have a "call server" button. And when they do, it often doesn't work, because the other end isn't hooked up to something that gets people's attention.
The much lower-tech and more enjoyable solution I've seen used successfully many many times is a call button on the table. Press the button if you want service. Bonus if it has a "press this other button if you want to be left alone".
I always wondered how actually cheaper a QR menu is, compared for instance to an android tablet you set there or pass to the client so they order.
People immediately see the cost of a tablet, the issues with charging, building the app and updating the data, customers not understanding the interface etc.
Yet, they don't seem to grasp all of these are basically the same for a QR code: you're now maintaining an online service just for your restaurant. People still need to understand your site. You'll be dealing with their browser, extensions, languages etc. And you still need the QR code somewhere on the table, visible, explain that it needs to be scanned (by one person ? everyone ? do they need to keep the tab open ? etc.). These will all be costs, you pay either upfront by hiring experts to make the experience good, or on the go with crappy customer experience, your staff spending more time than needed working around the system and explaining it to your customers, and those not coming back.
There are super cheap and simplistic ways to deal with all of that, and I assume GP has seen many of these.
PS: The weirdest experience I ever had was someone at our table scanning the code, to copy/paste it on the browser on their personal profile (default was work browser).
They missed a digit in the url params and ended up opening the order menu of a table at the other side of the restaurant, and just started ordering, then waiting for a pretty long time while the other table was dealing with the mess.
Notice that when someone grabs their phone, they're not just grabbing a web browser but also their personal entertainer. Besides chats, many people also have various apps trying to nudge them into engaging with it. I have never been in a restaurant without menu though, I just imagine that I'd end up on HN while contemplating what to pick out of sheer habit and muscle memory
Besides that, there are people who have no smartphone or an empty battery, can't use a smartphone (one of my grandparents can read a menu with a pocket magnifier, but a touchscreen phone is too cumbersome and so they don't know how to use that; another grandparent has memory issues and never learned how to use a smartphone), don't have signal (you need to provide WiFi), have an allergy and need to ask questions or ensure that some ingredient is really taken out, etc. I'm not against the option to order impersonally but a good kitchen communication system would go a lot deeper than slapping wetware with slopware
For whatever random reason, all the QR code menus where I live are implemented in a really, really shitty way: unresponsive, with huge latency, making it unnecessarily hard to make a choice. And good luck asking for a modification, like no mashed potatoes please or anything like that.
> ... without having to flag down a server / waiter.
Using QR codes to not having to talk to a human and to "optimize" humans out of the loop in a restaurant sounds like a great scenario for an episode of black mirror.
QR code menus are one the shittiest experience people going to a restaurant to enjoy a meal can have.
> Do you not like this option?
The entire world fucking hates it. It's totally dumbfucktarded.
Nice writeup but use legacy phone is really need to do some tuning and hacks, but that's fun to do if one like it.
For me I have a 24x7 running intel gen 11 nuc, i am lucky to get it installed with 32G memory and two NVME disks years ago and cannot image add those memory and NVME disk with much higher price
Just run some docker and VM on that, use cloudflare tunnel or tailscale funnel if you want to expose some service to the wild internet
I recently even find a pixel 5 and passthrough it to the openclaw VM on that nuc, let it browse the X to find some posts and relax me from doom scrolling, though seems timeline still bad and no much valued post find.
I think you are doing much better with an optiplex 3050 micro and libreboot. Android with it's locked down bootloader and outdated kernels is a dead platform. You are fighting Google's control at every step.
Sad, the US used to be all about freedom...
I thought about this a few times but the flash storage is a bit sketchy for running a database, I am worried it can die. Tho android does have sqlite built in and works fine.
Just because it's a database doesn't mean it does more write cycles on the flash cells than storing regular files would cause it to. Anytime you change something, it writes either a file, or does an update/insert query which causes a file write. In both cases you've written a few bytes to disk, doesn't really matter what the underlying system is (not majorly anyway, surely one solution will turn out to be some % better than another and cause the average phone's storage chip to last 30 years instead of 40 under some defined typical usage pattern)
You can also just insert an SD card to do the write-heavy operations and swap it out when it wears out (you make backups, right? :D)
Any storage device can die, that's why you have a robust backup system, right? Right...? :)
But the flash storage in a phone is going to be the eMMC variety which typically means a proper controller and TLC NAND. Not as robust probably as a "proper" SSD/NVME, but a lot more durable than most micro SD cards. And it'll have health indicators.
That's a test of the software more than the hardware. I run a lot of stuff on low-powered hardware and the main problems are when the software allocates a ton of memory unnecessarily (Java software is usually the worst offender) or does a million things for each request that are either unnecessary or duplicate.
Any efficient blog software will handle HN fine on anything that you can call a smartphone, so like maybe not a Nokia with GPRS and 6MB on-board storage, but something like the Samsung S2 from 2011 that basically shipped as a Linux box -- just need to adb push ARM binaries for PHP and a webserver
The main difference is that Camoufox doesn't have an iOS 6 client. Surf is built around that client experience: omnibox, tabs, history, uploads/downloads, touch input, and audio/video streaming targeting 60fps. So Camoufox is great for running browser instances on a VPS, but it doesn't really replace what Surf is doing for an old iPad.
> An open source anti-detect browser built for AI agents.
Sounds legit, circumventing the locks people are trying to add to their properties in an attempt to keep costs down and the service available for people that are actually trying to use the service rather than bots roaming around. Really happy for you that you managed to run ten of those. Not a nuisance at all.
That would be a fair point, but I don't see that in the source. The article's first mention of Surf is a link to a previous article <https://seg6.space/posts/resurrecting-ipad-mini/> where the motivation is mentioned:
> I wanted to turn [the old tablet] into a smooth, distraction free reading device, so I [got root and] transferred a few books with iFile and had exactly what I wanted [...] until the reading workflow left the book. \n\n Reading is not really offline anymore. Finding articles, checking references, downloading PDFs, opening links, and asking an LLM about something confusing all happen on the web. [...] [The ancient Safari engine can't provide] JavaScript, CSS, media, and browser APIs expected by current websites [and Apple didn't allow third-party engines.] [...] Since I was on vacation, this also felt like the perfect pet project to hand to an LLM and see how far it could get. Out of that experiment, Surf was born.
The point seems to have been getting iOS 6 to run a modern browser, which was experimentally built with an LLM, not so much have LLMs roam and masquerade themselves on the web via it
"My phone is a server now"
would've been better
--
> built in battery backup
Isn't it always recommended to remove battery if you plan to use a phone or a laptop as a server? ( Fire hazard )
or is it fine to just limit the charging to 80%?
I prefer to assume he's lugging a 12U mini relay rack around with him, including a UPS and cellular fallback for networking, so he can run an Android VM and use it as his cell phone.
See also, this atomic clock wristwatch: http://leapsecond.com/pages/atomic-bill/
"My phone is a server now" - OP's choice is a subtle one.
By saying "my server is now ..." implies a change of server platform to a phone but not where it came from.
Your offering: "my phone is a server now" implies a change of phone usage but does not indicate where the server came from.
The article itself tells all.
To me, "my server is now..." implies there was some server, it used to do server things, but now it has been repurposed to do something else.
When I read the title, I literally thought "wow, I've seen a lot of people make servers out of old unused phones, but this is the first time I've heard someone make a phone out of an old unused server!"
Agreed, a more appropriate title might be “my server now runs on a phone”
> By saying "my server is now ..." implies a change
Except the original title is "my server is a phone now", not "my server is now...".
I meant it from the server’s perspective: VPS before, phone now. The article clears up any ambiguity pretty quickly.
Without reading the article, the original title made me think it was going to be related to voip or telephony somehow. Not using a phone as a server.
Some newer phones support bypass charging, which powers the phone directly instead of continuously cycling the battery. Mine doesn't support that, unfortunately, so I've limited charging to 80% to reduce battery stress.
That's surprising to me. It's been a standard feature on laptops for 20 years.
Edit: Or I should say, it was 20 years ago. Maybe modern stuff is worse, but my M2 MacBook Air and Framework 13 Intel Core both power on without a battery.
> or is it fine to just limit the charging to 80%
LiPo batteries will fail over time if left fully charged. If you don't need the battery it's best to keep it somewhere around 40%. If 80% is the only option, that's usually good enough.
The primary failure mode starts when the cells are charged near their maximum voltage and kept there.
Afaik mobile electronics’ power circuitry is designed so that power goes straight from charging port to the processor, without constantly charging/discharging the battery, once charging is complete. Unless the system suddenly draws a lot of power (more than the charger is supplying) - which is unlikely in a smartphone.
This is true but lithium-ion batteries degrade even while not being charged or discharged, and this is worse at the extremes of the state of charge (including fully charged), and at higher temperature (and often a plugged in device is on constantly and therefore spends more time at a high temperature than it normally would), so it's still not great for the battery. A lot of devices have a mode where the batter is kept at a lower state of charge if it's been plugged in for a while to try to reduce this.
Some phones bypass the battery when plugged in, but not all do. I know it's advertised as a feature of some gaming phones because graphics-intensive gaming is one use case where you'd want to use the phone heavily while plugged in.
Even then, some phones that advertise bypass charging are actually using a software-based implementation, and they can still end up with swollen batteries after being plugged in for extended periods.
Steamdeck lets you set your preferred % and the dock does pass through charging/power, which I love because that combo guarantees the longest possible life for the battery. I turn it back to 100% for flights, otherwise leave it as 80% at all times. Mine is almost 4 years old now but I don’t feel even a hint of battery degredation yet.
That's what I thought when I read this. It will probably destroy this battery very quickly.
But in other hand, it is probably a pain in the ass to remove the battery and do some workarounds to make sure the phone correctly behaves without it.
> It will probably destroy this battery very quickly.
'Very quickly' seems to be measured in decades from what I see
Most people I know have their laptop docked, either near-permanently at home or taking it between home and a desk with dock (so it's nearly always plugged in when in use)
The 'servers' I've had were all old laptops - yes, with battery (don't have to buy a separate UPS if one is already included). I've now also had two non-server laptops that were plugged into power pretty much all the time. I treated my first laptop perfectly: taking out the battery anytime I'd be in a classroom for several hours and it was already full enough (you know, back when sliding out your battery was possible in the first place), but I'm not noticing any faster degradation on the systems I treat poorly. I'll activate the 80% cap function where it exists (that is, in the latest phones and my work laptop's UEFI config panel), but the evidence of it being helpful has yet to materialize
I have read that studies show that charging to 100% causes degradation and I trust those more than random people's anecdotes (and I'm a random person), but I'm starting to wonder if I've maybe just misunderstood the study results. Maybe it's like how smoking increases your chance for some cancers but you might also be fine in the end. I see people with plenty of much younger and better-treated batteries that can't sustain literally 30 seconds of use whereas mine are reasonably close to the new state (some haven't had many charge cycles, just sat at a stressy high voltage all the time, which was indeed supposed to destroy them), and I don't know how to explain that otherwise
What you can do, with some extra steps, is have Home Assistant installed (and have a Home Assistant server, I guess!) on the phone and use a smart power switch that's also associated with Home Assistant.
The phone's battery can be reported to Home Assistant, where an automation can be triggered to turn off power to the phone when the battery gets above a certain percentage and turned back on when it goes below.
Not the best solution, but certainly an alternative from messing around with trying to get the phone to work without a battery.
Couldn't you just have it charging on a time-regulated supply e.g. smart plug?
I thought the article would be about something VoIP-related.
Most phones won't boot without the battery. Obvious workaround is to connect a power supply (and large cap) directly to the BMS, but in my experience most phones will shut themselves down when they think they've drawn more than the capacity of their battery, even if the voltage is still high. You can sometimes get around that by rebooting the device every couple of hours (and there might be a software/firmware fix in some cases?), but it's a lot less annoying to just accept a small risk of a house fire.
You could throw the phone in an ammo canister, if the wifi AP is close enough. When I was building hobby quadcopters, which requires several ~half-pound batteries that are a chemistry optimized for maximum power density, a lot of guys recommended steel ammo cans for battery storage.
I’m also an ammo-case-for-drone-batteries guy but they aren’t running and don’t generate heat. A phone would cook right?
Depends on how well the canister insulates it. Steel conducts heat fairly well; it'd get hot inside but I'd guess some airflow around the canister would probably keep it from getting hot enough to cook the electronics. (definitely don't take my word for it lol)
Phones are very low power, likely just a slightly warm spot on the can that radiates, conducts, and convects the heat out. You could also use one of those peltier module phone coolers, to ensure the heat gets out of the phone and into the air faster.
in a steady state closed container a peltier cooler will just increase total heat and not help your phone.
> but it's a lot less annoying to just accept a small risk of a house fire.
Agree! But do get one or two of those "small automatic fire extinguisher" or whatever they are, that you can attach on stuff, for electrical panels and what not, and put right next to it together with a traditional fire alarm, and you're good to go :)
No extinguisher can stop a lithium battery fire.
Putting the phone into a metal box, and the box, onto something not flammable, would help.
Suspend metal box in larger metal box?
How many phones have you tested this with?
FYI you can't do this on phones with locked bootloaders as well as the author did. you could use termux but if termux runs on anything lower than android 8 termux will be significantly limited and you will need postmarketos and maybe fix support for it. with termux, at best things will be slow without root, as the author mentioned rooting resulted in a speed increase.
without root, you also can't bind to ports on any android phone, and rooting often requires an unlocked bootloader and not all android phones support bootloader unlocking. You also need an unlocked bootloader to install postmarket os.
so without binding to ports you can't run a web server unless you run it through a proxy server which the author did, or containers at native speed on postmarket os because you will not be able to flash the phone with postmarket os if you have missing kernel syscalls on android nor root the phone for the speed increase on android depending on your android kernel version i assume.
so with android containers may or may not work without emulating x86, and without root chances are it will be slow specially if you are forced to emulate it as x86 because it turns out you have missing syscalls, nor root to make them faster if they exist on android. again no chances to install postmarket os nor rooting to make containers faster, without bootloader unlocking.
not all android phones can have their bootloaders unlocked unless you find an exploit. https://github.com/zenfyrdev/bootloader-unlock-wall-of-shame
cve-2026-0047 if the phone is before March 2026
I mean root In android
I still think the best bang for your buck is an old desktop PC. Something like an i5 with 8GB RAM and an SSD will suit 99% of home server uses with capacity to spare.
Even in this environment of inflated hardware prices, you can get one for a reasonable price.
Although I do see the appeal in getting unconventional hardware to do what you want, just for the love of the game.
It only makes sense if we are measuring raw capability per dollar. Once you factor in noise, physical size, idle power, battery etc, I think the calculation changes.
For a home sever that's mostly idle, it is almost like commuting in a pickup truck.
Modern dell desktops (like 8th gen+) idle around 10W. Raspberry pi isn’t saving much
I’m impressed with that.
I had a Nuc 9 extreme with a 10gb SFP card in it.
It averaged 69w. The Mac Mini I replaced it with uses just under 5w.
A Mini is possible the worst bang for buck. Also, the horror that that is Desktop Docker can’t be over emphasised. I know there are better ways of using Docker on a Mac, but having got something working I can’t face rebuilding it.
also factoring in spare hardware. most people have more spare phones than PCs.
My server is an extremely quiet Xeon workstation: unlike 1U servers, workstation are typically quiet. The good thing with a Xeon workstation is that it supports ECC RAM only and there are plenty used. For there's no market for used Xeon systems, so you can find stuff like 22-cores / 44-threads CPUs for $25 or something.
Some mod Xeon systems to use regular PC PSU like Be Quiet! ones and Noctua fans, etc. but my HP Z440 workstation is so quiet already I didn't need to.
In my case it's an headless server and is sits between the vacuum cleaner and the cat's food: there are literally only two cables connected to it, the power cable and the ethernet cable.
I can live with the footprint of one tower.
I run many VMs and containers inside those VMs: Git server, pastebin, NAS, containerized/VM-isolated AI agents, etc. but I don't need it to be up 24/7. I basically only need those when I'm working. So electricity usage ain't a big concern to me as it's more like running 8/7.
I don't feel like I'm commuting in a pickup truck: I feel like I have the perfect machine for the task.
I just bought a refurbished Lenovo k14 with 32gb of ram and Corei7-1165G7 with 500gb for 350$ which I'll use as my 2nd home server.
My first home server is a 2nd hand gaming PC that I bought for 500$ with 32gb and 7700rx video card that I use for GPU based tasks but it's quite audible when it's working so wife and I agreed it has no place in the living room anymore and is being deported to the corner of my office.
You can still find really good bargains for home servers from refurbished good computers, and given how well solved tech reliability is now, it's really not worth buying new.
This is my situation exactly. Got an i5-4570 with 16GB, no GPU, running Unraid with a bunch of containers for Jellyfin, qBittorrent, Crafty Controller (minecraft), HomeAssistant, etc.
I'm using a PCIe expansion card to get double NVMe drives for the cache disk, and then three (up to five) 3.5" spinning 8TB drives for long term.
I tried really hard to get it on 32GB but sadly the motherboard only has two slots and only supports 8GB DDR3 modules. I literally ordered some of the rare late-generation 16GB ones and it refused to boot, with no firmware updates available from the vendor (Advantech, board is an AIMB-274).
> I tried really hard to get it on 32GB but sadly the motherboard only has two slots and only supports 8GB DDR3 modules.
Buy a refurbished Xeon workstation with DDR4 RAM. If you're in the US or China they're basically free and if you're in the EU, they aren't too pricey (like you can definitely find ones for less than 200 EUR). Slap your double NVMe expansion card, your two NVMEs and your 3 3.5" spinning rust in it. If the workstation doesn't come with a Xeon with lots of cores, just buy one on eBay for $25.
Yep. My 2010 server does everything I need incredibly well, and I see no point in retiring it anytime soon.
ASRock H67M-ITX, i3 2.5GHz, 16GB DDR3 1333Mhz, 2TB+4TB SSDs, DVD-RW, in this Mini-ITX Antec case[1].
[1]: https://www.antec.com/product/case/isk300-150
What's the idle draw on that?
I don't have a measured plug on it, but estimate roughly around ~32-38w.
Yeah ‘old phones’ are pretty capable devices.
I’ve wondered for a while if they could be used as an extension to your laptop like the old 80s PCs where you would plug in hardware.
So ‘make it a server’ is the obvious way I guess, and then you could do things like render webpages on there and then stream them as images to your actual computer (like the now shutdown Mighty app), or run ffmpeg on there to convert a video, etc
I've been saving phones in my drawer for years, waiting to combine them to form a "supercomputer".
Any day now!
A couple of old phones and a gaming laptop with a broken screen makes a beast of a cluster
Cool project, but slightly better: before the rampocalypse I bought one of those aliexpress fanless linux boxes designed to be a router. I put a 32GB stick of ram in it and a 250GB SSD. I run opnsense in a virtual machine to serve as my home router and then I put the rest of my home server stuff on other virtual machines on that same box. It's still fanless/silent. It takes up zero additional space because it is my router. It would already be running 24/7 because it is my router so the additional energy usage of an additional VM is likely minuscule. The latency to my router is basically zero since it is on the same physical box. And it is using a real SSD so it has faster storage with almost certainly better endurance.
Router with fanless gpu for AI load , nvme slot will be game changing
Love the idea, great work!
Those saying it's said backwards clearly aren't self-hosters!
iPhone hardware wise is like an amazing alternative to rpi (way better hardware)
But the software makes no sense at all for a server, with screen locking due to inactivity and whatnot, not to mention energy and network priority is designed around a consumer use case on the move.
If you run Linux on it and take advantage from the Apple hardware like the light sensors and accelerometer - maybe it makes sense for measuring wave height, sky color, attached to a buoy or oil rig or something, you can even show it on a map due to its GPS. Assuming the oil rig has WiFi or a cell phone booster.
Idk though, maybe for something that important you just get the right stuff - cool hobby nonetheless.
The main argument you give for it not making sense can easily be changed in Settings. The energy and network considerations have no impact on running a virtual machine in UTM – iPhones will give everything they can as long as thermals are fine.
I thought this was going to be a rant about QR code menus.
Can you please tell me more? I'm trying to build a self hosted application (with Claude) so people can scan a QR code and place orders / send messages to the kitchen without having to flag down a server / waiter. It is supposed to save time and money for the restaurant and for the patrons and customers it is supposed to be more convenient. Do you not like this option?
Every system like that I've ever used has three fundamental problems:
1) It's optimized for the restaurant first, and the patrons second or not at all, and it shows.
2) Someone thinks "hey, we could sell ads on this", and now it's even more a piece of garbage. (This is also what happened with all the little kiosk-tablets on the tables at some restaurants. We just set them face-down on their screen and talk to a human.)
3) Most of them never anticipate unexpected problems. They might have an "order X" button, but they often don't even have a "call server" button. And when they do, it often doesn't work, because the other end isn't hooked up to something that gets people's attention.
The much lower-tech and more enjoyable solution I've seen used successfully many many times is a call button on the table. Press the button if you want service. Bonus if it has a "press this other button if you want to be left alone".
I always wondered how actually cheaper a QR menu is, compared for instance to an android tablet you set there or pass to the client so they order.
People immediately see the cost of a tablet, the issues with charging, building the app and updating the data, customers not understanding the interface etc.
Yet, they don't seem to grasp all of these are basically the same for a QR code: you're now maintaining an online service just for your restaurant. People still need to understand your site. You'll be dealing with their browser, extensions, languages etc. And you still need the QR code somewhere on the table, visible, explain that it needs to be scanned (by one person ? everyone ? do they need to keep the tab open ? etc.). These will all be costs, you pay either upfront by hiring experts to make the experience good, or on the go with crappy customer experience, your staff spending more time than needed working around the system and explaining it to your customers, and those not coming back.
There are super cheap and simplistic ways to deal with all of that, and I assume GP has seen many of these.
PS: The weirdest experience I ever had was someone at our table scanning the code, to copy/paste it on the browser on their personal profile (default was work browser). They missed a digit in the url params and ended up opening the order menu of a table at the other side of the restaurant, and just started ordering, then waiting for a pretty long time while the other table was dealing with the mess.
Notice that when someone grabs their phone, they're not just grabbing a web browser but also their personal entertainer. Besides chats, many people also have various apps trying to nudge them into engaging with it. I have never been in a restaurant without menu though, I just imagine that I'd end up on HN while contemplating what to pick out of sheer habit and muscle memory
Besides that, there are people who have no smartphone or an empty battery, can't use a smartphone (one of my grandparents can read a menu with a pocket magnifier, but a touchscreen phone is too cumbersome and so they don't know how to use that; another grandparent has memory issues and never learned how to use a smartphone), don't have signal (you need to provide WiFi), have an allergy and need to ask questions or ensure that some ingredient is really taken out, etc. I'm not against the option to order impersonally but a good kitchen communication system would go a lot deeper than slapping wetware with slopware
I don't think anyone over 30 around me likes that. At that point it's very close to door dashing but less convenient.
For whatever random reason, all the QR code menus where I live are implemented in a really, really shitty way: unresponsive, with huge latency, making it unnecessarily hard to make a choice. And good luck asking for a modification, like no mashed potatoes please or anything like that.
Not when I forget my phone in the car.
Having backup tablets on hand is a good hybrid.
> ... without having to flag down a server / waiter.
Using QR codes to not having to talk to a human and to "optimize" humans out of the loop in a restaurant sounds like a great scenario for an episode of black mirror.
QR code menus are one the shittiest experience people going to a restaurant to enjoy a meal can have.
> Do you not like this option?
The entire world fucking hates it. It's totally dumbfucktarded.
Sorry to be blunt.
Nice writeup but use legacy phone is really need to do some tuning and hacks, but that's fun to do if one like it.
For me I have a 24x7 running intel gen 11 nuc, i am lucky to get it installed with 32G memory and two NVME disks years ago and cannot image add those memory and NVME disk with much higher price
Just run some docker and VM on that, use cloudflare tunnel or tailscale funnel if you want to expose some service to the wild internet
I recently even find a pixel 5 and passthrough it to the openclaw VM on that nuc, let it browse the X to find some posts and relax me from doom scrolling, though seems timeline still bad and no much valued post find.
I think you are doing much better with an optiplex 3050 micro and libreboot. Android with it's locked down bootloader and outdated kernels is a dead platform. You are fighting Google's control at every step. Sad, the US used to be all about freedom...
I thought about this a few times but the flash storage is a bit sketchy for running a database, I am worried it can die. Tho android does have sqlite built in and works fine.
Just because it's a database doesn't mean it does more write cycles on the flash cells than storing regular files would cause it to. Anytime you change something, it writes either a file, or does an update/insert query which causes a file write. In both cases you've written a few bytes to disk, doesn't really matter what the underlying system is (not majorly anyway, surely one solution will turn out to be some % better than another and cause the average phone's storage chip to last 30 years instead of 40 under some defined typical usage pattern)
You can also just insert an SD card to do the write-heavy operations and swap it out when it wears out (you make backups, right? :D)
Any storage device can die, that's why you have a robust backup system, right? Right...? :)
But the flash storage in a phone is going to be the eMMC variety which typically means a proper controller and TLC NAND. Not as robust probably as a "proper" SSD/NVME, but a lot more durable than most micro SD cards. And it'll have health indicators.
I did a similar thing recently.
I run kube in a VM, and it was ressource constrained.
I hooked an old samsung to my pc and have let claude handle everything.
I only had to unlock the bootloader, flash the image claude uploaded.
It rebuilt postmarketos multiples times to enable various flag, and now there is a kube node running on it.
did you commit or open source your changes to postmarket os? also you can't unlock bootloaders anymore on samsung phones. you will need to find an exploit. https://github.com/zenfyrdev/bootloader-unlock-wall-of-shame
Claude did the changes, I didn't do anything, I think it's was mostly recompiling with flags.
I plan to do this when I buy a new phone, thank you for the great writeup!
Would have been a fun load test had you hosted your blog on the phone.
That's a test of the software more than the hardware. I run a lot of stuff on low-powered hardware and the main problems are when the software allocates a ton of memory unnecessarily (Java software is usually the worst offender) or does a million things for each request that are either unnecessary or duplicate.
Any efficient blog software will handle HN fine on anything that you can call a smartphone, so like maybe not a Nokia with GPRS and 6MB on-board storage, but something like the Samsung S2 from 2011 that basically shipped as a Linux box -- just need to adb push ARM binaries for PHP and a webserver
Shouldn't it be "My phone is a server now"?
I meant it from the server’s perspective: VPS before, phone now. The article clears up any ambiguity pretty quickly.
but this way is provocative, it gets the people going
“My daily driver is a clown car now.”
I've never heard of Surf but I can run 10x camoufox instances on a $8 VPS no problem. I highly recommend it.
The main difference is that Camoufox doesn't have an iOS 6 client. Surf is built around that client experience: omnibox, tabs, history, uploads/downloads, touch input, and audio/video streaming targeting 60fps. So Camoufox is great for running browser instances on a VPS, but it doesn't really replace what Surf is doing for an old iPad.
Not having heard of that, either, I looked it up:
> An open source anti-detect browser built for AI agents.
Sounds legit, circumventing the locks people are trying to add to their properties in an attempt to keep costs down and the service available for people that are actually trying to use the service rather than bots roaming around. Really happy for you that you managed to run ten of those. Not a nuisance at all.
It's in the same category as Surf, which is the pain point in the article. It's relevant.
That would be a fair point, but I don't see that in the source. The article's first mention of Surf is a link to a previous article <https://seg6.space/posts/resurrecting-ipad-mini/> where the motivation is mentioned:
> I wanted to turn [the old tablet] into a smooth, distraction free reading device, so I [got root and] transferred a few books with iFile and had exactly what I wanted [...] until the reading workflow left the book. \n\n Reading is not really offline anymore. Finding articles, checking references, downloading PDFs, opening links, and asking an LLM about something confusing all happen on the web. [...] [The ancient Safari engine can't provide] JavaScript, CSS, media, and browser APIs expected by current websites [and Apple didn't allow third-party engines.] [...] Since I was on vacation, this also felt like the perfect pet project to hand to an LLM and see how far it could get. Out of that experiment, Surf was born.
The point seems to have been getting iOS 6 to run a modern browser, which was experimentally built with an LLM, not so much have LLMs roam and masquerade themselves on the web via it
keeping phones connected to power 24/7 is a fire hazard
Dunno, i ran my blog for two years from my old iphone 4. It was fine.
Doesn't this depend on the kind of phone and battery?
Keeping fans connected to power 24/7 is a fire hazard, too.
Running electricity through wood, arguably insane.