Things you made that could be valuable to other people if they used it, but which you abandoned because you enjoyed making the thing more than you would have enjoyed marketing it.
I built a managed Kubernetes platform because it was what I knew how to build. Pretty quickly I realized selling it was actually going to be pretty similar to my day job which is what I was seeking a change from.
So I took a left turn and started making a VR game instead. The marketing path for that is much more interesting.
The data is small enough that you can keep it all in memory. It was a nice toy and it let me play with a few data retrieval techniques I had heard talking about but never actually implemented.
The data is scraped from restaurant menu aggregators. There are some online and they don't seem to even try to stop you from scraping.
If I had to guess I'd say that scraping restaurant websites would be a much bigger project than what I did, given that there isn't structure in how menus are presented, and most of them still have pdfs, or even worse, an image for their menu.
This is sweet haha. I want Mapo Tofu, nearby, ranked by price and distance.
Next: what's the sentiment about that dish (from reviews). Or at least a restaurant rating.
Love how fast it is and the non distracting design. Bookmarked.
I have built an iOS and Android app called Boomerang to send email to myself in one tap. It’s been around for 5 years and has started as a side project for learning Ionic. I have maintained it so far and users (5K+ MAU) love it so far and want to give me money for it. I am planning to monetize it though as its started to cost me some money.
I'm building a tool to help people not give up on marketing (something I've struggled with in the past). If you're interested in trying it out please email me. Email on my profile.
I built my own GeForce Now / Google Stadia clone for up to 8 gamepads to connect to the same instance. It's GPU-accelerated, UDP-based and all C++.
It works well and I still use it with friends and family to play games together. I even designed a logo, registered it for an EU trademark, and made t-shirts.
But when Steam Remote Play Together was announced, I decided that there wasn't a market for it anymore, so instead I switched to building real-time AI tools in C++.
My next idea, funny Webcam filters for desktop, was then pre-empted by Snap Camera coming to desktop. With so many FAANG-scale companies releasing free tools, it has become really tricky to find financially viable niches.
When a deeply entrenched player gets into the game, small players do not stand much of a chance. I think few people have the necessary David in them to go up against such Goliaths. OP made the personal choice to walk away from an uphill battle, and I don’t blame them one bit.
Yes, the market surely exists. But I can't build a bootstrapped business around competing with FREE. Steam has other ways to monetize user engagement, but I would have needed to charge users for the service.
Maybe a similar project could take off: sharing console instances. Flat community of 50 apartments would buy 5 consoles and share them with each other (you keep them all in a storage room and just connect with TV via ethernet).
One major problem is that console manufacturers would probably need to be a part of this and I think they would rather sell 10 consoles individually than 5 for sharing.
> One major problem is that console manufacturers would probably need to be a part of this and I think they would rather sell 10 consoles individually than 5 for sharing.
It would certainly work better with manufacturer involvement, but you could use HDMI capture to source the video, and USB to provide the inputs, so the console doesn't know any better. There's some licensing rigamorole and content management as well. You could probably do a forced log out between users and rely on cloud saves, but that's kind of ugh... Would work better if you can trust the people you share the console(s) with.
Yes, in principle you could do that (apart from all the legal troubles...)
In practice we like things customized and only want to see our own savegames available on the list so this is a real issue that could break this especially if large number of strangers were to use it.
Another issue is waste - what if you want to play some game but it was installed on another console that is currently playing something completely different? Installing all games on all consoles seems crazy. Maybe have a large NAS with multiple games preinstalled? If we could add another type of game licensing which would allow for playing the same game on 5 different hardware consoles simultaneously that would be perfect.
Resources would be used to their maximum potential.
Scheduling around game installs would be tricky yeah. I don't know how the various console do with having their storage swapped... I know the switch can store downloaded games to SD cards, but I don't know if it lets you put that SD card into another console and play the games if a licensed user is logged in... Or if it needs to be downloaded by that specific console. I think the current xbox and playstations also have user replacable storage, but again, it may be specialized per console? I know the platforms already have code to check for licensed user on a downloaded game when the game is downloaded to a console not deemed the 'home' console for a user.
If you can swap storage, you can build around NAS connected storage emulators.
Otherwise, yeah, big storage and guessing about what should go where. Like Netflix CDN nodes, but with more black boxes.
How much does it cost for a GPU accelerated VM in the cloud? It would be great if there was an automated way to spin up a VM, install your software and then destroy the VM when you're done. After all, GeForce Now and Google Play don't let you play your full Steam library.
The NVIDIA T4 VWS GPUs aren't as flashy as they used to be and there's no option for faster CPU cores in GCP, but for medium-setting ray traced gaming it'll still do the trick!
This interesting for me from the standpoint of shared controllers. What is the general idea behind this? Did you write a driver that would fake these controllers on the 'server'? What platforms have you covered? I'm into massively local multiplayer game and was always toying with ideas on how to reliably connect as much controllers as possible into a game.
So each client polls the local controller using the DX APIs, streams the state to the server and then the server creates virtual gamepads to send the input to the game.
A day-pass VPN with a single server, with SMS capabilities (with your GSM SIM card). Sort of like a VPN vending machine.
Tech-stack: OpenVPN + MySQL + some Python scripts as glue + PayPal integration.
Use case: If you need a Venezuelan IP address [0]. you'd just access the VPN front site, pay for a day-pass, and use the VPN [1]. Also banks use SMS-based 2FA, so having the SIM card in the server helps [2].
Turns out there are cheaper VPNs that do have Venezuelan IP addresses, though my offered benefit is to offer home random IP addresses, which makes it more stealth. Also the economics didn't help, and the upstream bandwidth isn't great as well.
I took down the site and just left the VPN for myself and some trusted friends. Also, if someone did something sketchy with it, it'd by tied to my home address.
Maybe someday I'll create a blog post about it.
--
[0]: Something very niche, banks here don't allow people foreign IP addresses to do bank transfers, etc., they could also block your bank account entirely.
[1]: It'd issue short lived certificates for the amount of time purchased.
Based on the URL, I thought it was going to be the same thing that I had built. It was a simple web app for the people who lend and/or borrowed a book or a CD or something from a friend, but after too long don't remember which one. While still fresh in mind, one would enter:
I built https://www.quikk.co.uk which gives users a better audit trail and backup of their Xero accounting data. I’ve never really marketed it, but MRR-wise it’s doing ok. Promising volumes so far. I keep meaning to put effort into getting it into the hands of those who would benefit most from it but always get sidetracked with my other projects.
Throwing this out there, why not… if anyone would like to help me, let me know!
I made a simple tool to extract comments from PDFs with different severity levels. Mainly developed for my use marking up documents for review in an academic context, a need many academics have but I'm not sure how many others.
I’m actively working on language learning software through immersion at https://polyglatte.com . It’s currently completely free with no way to pay for it but I still feel some amount of guilt about self-promotion even though I get lots of positive feedback and it’s already helping people.
I find marketing to be very uncomfortable so I’ve been neglecting it, but I know it’s the only way to any kind of success.
Not sure if this would be valuable to anyone but I have a custom apps script that parses all of my new calendar event and takes certain actions on certain events. For example, events from/to my wife get marked as personal color and private, title starts with a star -> personal and private. Event from workout class system -> mark as fitness and private.
I took a simple google ticket system tutorial and made a whole gapps-script based internal company ticket portal. Using googles auth system meant user-correlatiom became trivial and sped up task resolution. As I was the only person running the IT department, I had the ability to choose and I didn't like any of the corporate products for a company that did business more frugally than falling for death by a thousand license cuts, a very real danger.
Then gapps-script updated, broke some stuff, I got tired of maintaining it, and we got a new dev, so I aquiesced to Jira.
It is a very easy-to-use self-contained JS framework for prototyping and making stuff with Android devices.
You can access sensors, NFC, bluetooth, MQTT, websockets, or write a simple UI with a couple of lines of code.
Think in the terms of easiness of Arduino for Android
When people try it they always say WOW "why have I never heard of this?!" ( probably because I'm very bad at marketing :/ )
I got it mentioned in Hackaday and some comments here in HN but that's all :)
Looks pretty good. Is there anyway to share a completed prototype app with a other user, or run an app developed this way standalone? I went to look at the docs page on your site, but it's more of a basic tutorial rather than docs.
My Weight Loss Story 40kg in 40 Weeks (90lb in 9 months) Without Exercise [0]
I didn't necessarily enjoy writing it, or the process, but thought it needed to be written because I have heard so many different weight loss suggestions through out my life.
I made a tool to export YouTube subtitles directly into Anki without downloading the video, it must be useful since it even got me a job offer by one of the users (pretty famous person).
I never did more marketing than a Reddit post when the proof of concept was made.
Coincidentally, I've worked on it recently and added some useful functionality.
What made you build it? It looks like one would use it for learning a language... mapping written text to pronounciation on flashcards? Why do you think making your own cards from YouTube subs is better than dedicated language apps that already include spoken audio?
I was inspired by this tool, http://www.randomhacks.net/substudy/
It required to cut the video, with a lot of processing, cli knowledge and to storage the cut videos taking a lot of space, my idea was to make the process as lightweight as possible, the outcome of the extension it's a CSV with the marks for when the video starts and stops for each sentence, it has a negligible output size in comparison and can be used in low-end computers.
You described the functionality perfectly, I made it as a tool to have custom cards for practising languages without having to be locked in any platform, in my opinion language platforms are very bland with their spoken lessons and only have perfectly pronounced sentences, with the tool you can practice more real and engaging topics for yourself.
Another usage that some users commented was to use it as a learning tool for lectures and presentations, that's why I included a merge subtitles functionality recently, so the cards can hold more information as opposed a single sentence.
I was frustrated with how slow and clunky all the existing solutions were for regex searches. The built in file search in MSVC has a UI that wastes a tonne of space (and the search itself is unuseably slow). Constantly changing to a terminal to run ag/grep and then open the file manually in MSVC was a pain, so I learned how to build an extension and put a GUI onto the command line tools I was already using.
Tools like Visual Assist are great if you want actual symbol indexing, but the searches (of a UE4) codebase can take 30-60s. TurboSearch can run a regex search of the entire engine in ~6s and searches of just the game code take ~200ms.
I have only sold one copy, but I use it every day and it's a core part of my workflow. I thought there would be more interest in it when I built it, but since I've gotten so much use out of it myself, the time spent building it was well worth it.
A browser extension for Twitch.tv (mainly for MOBAs) that enabled live-stream viewers to mouse-over in-game items/abilities and be shown a tooltip describing that item/ability as if they were playing the game themselves. Useful for viewers who are new to watching competitive matches and have no idea what certain icons/abilities/items represent.
Scraped millions of publicly available property tax documents for a specific region and combined it with Leaflet and OpenStreetMap to visualize the data.
304 comments
[ 4.1 ms ] story [ 287 ms ] threadSo I took a left turn and started making a VR game instead. The marketing path for that is much more interesting.
The data is small enough that you can keep it all in memory. It was a nice toy and it let me play with a few data retrieval techniques I had heard talking about but never actually implemented.
If I had to guess I'd say that scraping restaurant websites would be a much bigger project than what I did, given that there isn't structure in how menus are presented, and most of them still have pdfs, or even worse, an image for their menu.
Love how fast it is and the non distracting design. Bookmarked.
https://boomerang-app.io
It works well and I still use it with friends and family to play games together. I even designed a logo, registered it for an EU trademark, and made t-shirts.
But when Steam Remote Play Together was announced, I decided that there wasn't a market for it anymore, so instead I switched to building real-time AI tools in C++.
My next idea, funny Webcam filters for desktop, was then pre-empted by Snap Camera coming to desktop. With so many FAANG-scale companies releasing free tools, it has become really tricky to find financially viable niches.
Maybe a similar project could take off: sharing console instances. Flat community of 50 apartments would buy 5 consoles and share them with each other (you keep them all in a storage room and just connect with TV via ethernet).
One major problem is that console manufacturers would probably need to be a part of this and I think they would rather sell 10 consoles individually than 5 for sharing.
It would certainly work better with manufacturer involvement, but you could use HDMI capture to source the video, and USB to provide the inputs, so the console doesn't know any better. There's some licensing rigamorole and content management as well. You could probably do a forced log out between users and rely on cloud saves, but that's kind of ugh... Would work better if you can trust the people you share the console(s) with.
In practice we like things customized and only want to see our own savegames available on the list so this is a real issue that could break this especially if large number of strangers were to use it.
Another issue is waste - what if you want to play some game but it was installed on another console that is currently playing something completely different? Installing all games on all consoles seems crazy. Maybe have a large NAS with multiple games preinstalled? If we could add another type of game licensing which would allow for playing the same game on 5 different hardware consoles simultaneously that would be perfect.
Resources would be used to their maximum potential.
If you can swap storage, you can build around NAS connected storage emulators.
Otherwise, yeah, big storage and guessing about what should go where. Like Netflix CDN nodes, but with more black boxes.
I did not yet try it myself (maybe it doesn't even work anymore?), and it might not be an option for businesses.
(haven't given it a try just yet but surely looking forward to)
The NVIDIA T4 VWS GPUs aren't as flashy as they used to be and there's no option for faster CPU cores in GCP, but for medium-setting ray traced gaming it'll still do the trick!
So each client polls the local controller using the DX APIs, streams the state to the server and then the server creates virtual gamepads to send the input to the game.
Tech-stack: OpenVPN + MySQL + some Python scripts as glue + PayPal integration.
Use case: If you need a Venezuelan IP address [0]. you'd just access the VPN front site, pay for a day-pass, and use the VPN [1]. Also banks use SMS-based 2FA, so having the SIM card in the server helps [2].
Turns out there are cheaper VPNs that do have Venezuelan IP addresses, though my offered benefit is to offer home random IP addresses, which makes it more stealth. Also the economics didn't help, and the upstream bandwidth isn't great as well.
I took down the site and just left the VPN for myself and some trusted friends. Also, if someone did something sketchy with it, it'd by tied to my home address.
Maybe someday I'll create a blog post about it.
--
[0]: Something very niche, banks here don't allow people foreign IP addresses to do bank transfers, etc., they could also block your bank account entirely.
[1]: It'd issue short lived certificates for the amount of time purchased.
[2]: With a ZTE MF667 USB modem.
- their email
- their friend's email
- the name of the thing being shared
- and the date both would be reminded (by email).
Throwing this out there, why not… if anyone would like to help me, let me know!
I don't even know if anybody else ever used it, as I don't track anything at all - but most likely not.
https://github.com/hoffmangroup/pdfcomments
Also a peer to peer parking spot rental platform a la Airbnb. JS with a python backend.
A gamified voter registration Facebook app. Rails backend.
Not that I didn’t think I would market them, at the time I built them I was just unaware of how to effectively, and I have newer projects since.
I find marketing to be very uncomfortable so I’ve been neglecting it, but I know it’s the only way to any kind of success.
Then gapps-script updated, broke some stuff, I got tired of maintaining it, and we got a new dev, so I aquiesced to Jira.
I hate jira.
It is a very easy-to-use self-contained JS framework for prototyping and making stuff with Android devices.
You can access sensors, NFC, bluetooth, MQTT, websockets, or write a simple UI with a couple of lines of code. Think in the terms of easiness of Arduino for Android
When people try it they always say WOW "why have I never heard of this?!" ( probably because I'm very bad at marketing :/ )
I got it mentioned in Hackaday and some comments here in HN but that's all :)
If the other person has PHONK installed you can just share the project folder.
I used to have an "Export/Import PHONK file" back in the days but it broke at some point and I decided to remove since nobody noticed :D
It has one user: me
I didn't necessarily enjoy writing it, or the process, but thought it needed to be written because I have heard so many different weight loss suggestions through out my life.
[0] http://40in40book.com
https://rate.house/ - like a collaborative IMDb but also has music, books, video games, and podcasts
https://wordhoot.com/ - competitive Wordle
None of them are abandoned but I just don't have the resources to do any marketing.
I never did more marketing than a Reddit post when the proof of concept was made. Coincidentally, I've worked on it recently and added some useful functionality.
Extension: https://chrome.google.com/webstore/detail/youtube2anki/boebb... Source: https://github.com/dobladov/youtube2Anki
You described the functionality perfectly, I made it as a tool to have custom cards for practising languages without having to be locked in any platform, in my opinion language platforms are very bland with their spoken lessons and only have perfectly pronounced sentences, with the tool you can practice more real and engaging topics for yourself.
Another usage that some users commented was to use it as a learning tool for lectures and presentations, that's why I included a merge subtitles functionality recently, so the cards can hold more information as opposed a single sentence.
I was frustrated with how slow and clunky all the existing solutions were for regex searches. The built in file search in MSVC has a UI that wastes a tonne of space (and the search itself is unuseably slow). Constantly changing to a terminal to run ag/grep and then open the file manually in MSVC was a pain, so I learned how to build an extension and put a GUI onto the command line tools I was already using.
Tools like Visual Assist are great if you want actual symbol indexing, but the searches (of a UE4) codebase can take 30-60s. TurboSearch can run a regex search of the entire engine in ~6s and searches of just the game code take ~200ms.
I have only sold one copy, but I use it every day and it's a core part of my workflow. I thought there would be more interest in it when I built it, but since I've gotten so much use out of it myself, the time spent building it was well worth it.
Scraped millions of publicly available property tax documents for a specific region and combined it with Leaflet and OpenStreetMap to visualize the data.