TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ask HN: Moving a not-for-profit web app off AWS

39 点作者 sjayasinghe4 个月前
I&#x27;m developing an app to help foreign Buddhist monks in Thailand learn Thai and Pali languages. The app incorporates LLMs for interactive learning, and I&#x27;m experimenting with spaced repetition techniques within chatbot interfaces and for long-form text memorization.<p>Right now I&#x27;m running everything on AWS, but the cost feels overkill for my current needs. Since my past professional background is in AWS, it was the default choice, but given that this project is currently funded through my own savings and I&#x27;m currently unemployed, I&#x27;m looking to minimize costs. Scalability isn&#x27;t a concern as the user base is small.<p>Currently I&#x27;m using 3rd party LLM APIs, but in the future I may want to explore serving open-source models that I fine-tune specifically for language learning and Buddhist text exploration.<p>Does anyone have recommendations for budget friendly alternatives that would suit these needs? Ideally, something that could handle a future transition to self-hosted models if needed.<p>Also, if anyone has experience with applying spaced repetition or other memorization techniques to chatbot interfaces and long-form text retention, I&#x27;d love to hear your thoughts.

29 条评论

isoprophlex4 个月前
Few things beats a cheap, powerful Hetzner server, IMO. I host a LOT of stuff on my single €40&#x2F;month box. 20 core, 64 gb ram, 1 TB ssd, unmetered network.<p>They have GPU-backed servers too, obviously theyre more expensive, and not ideal for your usecase maybe.<p>Modal is a good serverless alternative if you want to scale to zero and be able to handle spiky loads.
评论 #42822417 未加载
评论 #42826069 未加载
评论 #42822215 未加载
评论 #42822349 未加载
评论 #42823197 未加载
indigodaddy4 个月前
What&#x27;s your current resource usage? Memory&#x2F;CPU&#x2F;disk etc?<p>netcup.eu has instances that are generally about the best value you&#x27;ll find anywhere. They&#x27;ve been around a long time and good rep.<p>buyvm.net while not quite as cheap as a random lowend host (keep in mind there are lots of fly-by-night lowend hosts with crazy prices (buyvm is NOT one of those)), and harder to find VM vacancies there, is super solid, fantastic rep, great network and large transfer, if you need large block device attached they have up to 10TB &quot;slabs&quot; you can attach to your VM for $5&#x2F;TB&#x2F;mo (you won&#x27;t beat that pricing anywhere-- I&#x27;ve had one attached for a year or so and no issues very stable. And I definitely use the space). And they have a fantastic discord community and documentation&#x2F;help, and are fast to respond to issues in their Discord.<p>But in general lowendtalk.com offers section and lowendbox are your friend (but pay close attention to provider reviews&#x2F;offer comments to avoid unreliable and&#x2F;or fly-by-night hosts).
simonw4 个月前
What&#x27;s your technology stack?<p>If you have a Python or Node.js or similar there are a wealth of inexpensive options. I personally really like Fly because the pricing is predictable (you effectively pay per container per month) and it&#x27;s easy to get anything that runs in Docker to run there. <a href="https:&#x2F;&#x2F;fly.io&#x2F;docs&#x2F;about&#x2F;pricing&#x2F;">https:&#x2F;&#x2F;fly.io&#x2F;docs&#x2F;about&#x2F;pricing&#x2F;</a> starts at $1.94&#x2F;month for a 256MB container which may be enough if your backend is small and efficient.<p>Heroku is a bit more expensive but still a relatively cheap and very robust and proven option.<p>If you want to maximize the performance you can get for your money a dedicated server from Hetzner is great value, but you&#x27;ll have to do a lot more management than you would with a PaaS like Fly or Heroku.<p>Don&#x27;t spend too much time worrying about serving your own models. This is a MUCH harder problem - you need GPU allocations etc - and a whole lot more expensive. Fine-tuning models is rarely the right solution so I don&#x27;t think it&#x27;s worth assuming you&#x27;ll be doing that. If you DO do that you can solve that hosting problem separately later on.<p>(Fly have a GPU product which is great but like all GPUs it&#x27;s expensive compared to regular hosting.)
password43214 个月前
Though not intended for reliable, professional, and&#x2F;or commercial use, Oracle Always Free⁰ is free cloud VPS (24GB RAM [ARM], 200GB, 10TB transfer, IP) and Scaleway Stardust¹ is no-additional-cost data transfer capped at 100mbps (€3.43&#x2F;mo: 1GB RAM, 10GB, IP). Maybe it goes without saying, but don&#x27;t mess around on the info requested when creating an account - they won&#x27;t.<p>If you&#x27;re not interested in sysadmin work... well, that&#x27;s what you&#x27;re paying extra for! I&#x27;ve never bothered but you can find discussions of open source PaaS options here on HN keword Dokku².<p>I see no reason not to buy or repurpose a dedicated laptop&#x2F;Raspberry Pi&#x2F;NUC and host everything on your own home internet connection while getting started. This helps separate concerns a bit; less secure would be a VM (probably ok) or even docker container (yikes!). Cloudflare Tunnel³ is a free endpoint vs. NAT port forwarding exposing your home IP.<p>I can&#x27;t provide the voice of experience regarding self-hosted LLMs, but there have been a few HN discussions on building your own. You might find a deal on a near-dead GPU and can buy &#x27;without graphics card&#x27; gaming machines capable of powering them from others reselling after pulling new GPUs.<p>⁰<a href="https:&#x2F;&#x2F;www.oracle.com&#x2F;cloud&#x2F;free" rel="nofollow">https:&#x2F;&#x2F;www.oracle.com&#x2F;cloud&#x2F;free</a><p>¹<a href="https:&#x2F;&#x2F;www.scaleway.com&#x2F;en&#x2F;stardust-instances" rel="nofollow">https:&#x2F;&#x2F;www.scaleway.com&#x2F;en&#x2F;stardust-instances</a><p>²<a href="https:&#x2F;&#x2F;hn.algolia.com&#x2F;?query=dokku%20comments%3E0" rel="nofollow">https:&#x2F;&#x2F;hn.algolia.com&#x2F;?query=dokku%20comments%3E0</a><p>³<a href="https:&#x2F;&#x2F;developers.cloudflare.com&#x2F;cloudflare-one&#x2F;connections&#x2F;connect-networks" rel="nofollow">https:&#x2F;&#x2F;developers.cloudflare.com&#x2F;cloudflare-one&#x2F;connections...</a>
评论 #42823248 未加载
yoaviram4 个月前
If you form a real nonprofit AWS will give you up to 5000 USD per year.<p><a href="https:&#x2F;&#x2F;aws.amazon.com&#x2F;government-education&#x2F;nonprofits&#x2F;nonprofit-credit-program&#x2F;" rel="nofollow">https:&#x2F;&#x2F;aws.amazon.com&#x2F;government-education&#x2F;nonprofits&#x2F;nonpr...</a>
评论 #42822554 未加载
评论 #42822681 未加载
lunarcave4 个月前
Have you tried applying to AWS Activate? I&#x27;m not sure whether they support non-profits, but if you can swing it, they&#x27;ll subsidise your compute and some model costs.<p>In terms of a budget friendly alternatives, definitely look into fly.io. We have a AWS setup, but run some compute for ephemeral use cases in fly.io.<p>&gt; Also, if anyone has experience with applying spaced repetition or other memorization techniques to chatbot interfaces and long-form text retention, I&#x27;d love to hear your thoughts.<p>Have some experience with conversational experiences and LLMs. It all depends on what the modality of the interaction is. Is it something that users can have a long chat conversation with? Or is it more like the anki experience where &quot;flashcards&quot; get presented to you?
评论 #42829331 未加载
coderintherye4 个月前
If it&#x27;s a simple app, consider NearlyFreeSpeech <a href="https:&#x2F;&#x2F;www.nearlyfreespeech.net&#x2F;services&#x2F;hosting" rel="nofollow">https:&#x2F;&#x2F;www.nearlyfreespeech.net&#x2F;services&#x2F;hosting</a><p>Easy, straight-forward no-frills hosting.<p>Could you share the app if&#x2F;when it&#x27;s available somewhere? I have a Buddhist monk friend moving to Thailand later this year who would be quite interested in something such as what you described.
评论 #42823048 未加载
评论 #42829397 未加载
emmanueloga_4 个月前
Cloudflare is a nice option if you can adapt your system to run there [1]. You could also run just the frontend on CF and host any APIs elsewhere. Or check these sites [2] [3] with some good reviews and comparisons of hosting providers.<p>Regarding spaced repetition, RemNote 1.17 an 1.18 has a bunch of AI features for summarizing PDFs and YouTube transcripts and turning that into Q&amp;A for spaced repetition, <i>and</i> they also have a Chatbot interface in-app. Perhaps you could check some of what they are doing for inspiration [4].<p>--<p>1: <a href="https:&#x2F;&#x2F;developers.cloudflare.com&#x2F;workers&#x2F;languages&#x2F;" rel="nofollow">https:&#x2F;&#x2F;developers.cloudflare.com&#x2F;workers&#x2F;languages&#x2F;</a><p>2: <a href="https:&#x2F;&#x2F;getdeploying.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;getdeploying.com&#x2F;</a><p>3: <a href="https:&#x2F;&#x2F;www.vpsbenchmarks.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.vpsbenchmarks.com&#x2F;</a><p>4: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;@RemNote&#x2F;videos" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;@RemNote&#x2F;videos</a>
评论 #42823445 未加载
评论 #42829292 未加载
9999000009994 个月前
I can second Hetzner, I went from spending about $40 a month to $5 a month to host two blogs.<p>The ID verification seems to be a bit random, I didn&#x27;t have any issues with it. I would suggest looking at how much you&#x27;re actually spending per month, and then think about how much time and effort moving to another provider will be.<p>There&#x27;s a very good chance you&#x27;ll end up spending so much time moving this to another provider, you could have used that time to find work opportunities or do something else worthwhile. AWS credits are pretty easy to get, assuming you haven&#x27;t gotten them before.
sjayasinghe4 个月前
OP here with additional details requested by several of the commenters. Thanks so much to everyone for your responses on this thread, as well as those who reached out via email with suggestions and opportunities for collaboration.<p>Description of the stack:<p>- The frontend is a TypeScript Next.js app, with Prisma and RDS Postgres for persistence, and S3 for storing files uploaded by users.<p>- On the backend I have a BullMQ worker for handling asynchronous tasks such as audio&#x2F;video transcription, backed by Redis.<p>- A Telegram bot that allows quickly adding resources from a phone and for having practice conversations with AI.<p>- Everything is deployed via a Github Actions pipeline, with images hosted on ECR and other AWS resources deployed using Terraform.<p>Preferably the solution would allow me to reuse at least parts of the existing CI&#x2F;CD pipeline and Terraform for managing resources. Some of the suggestions by commenters such as running this on a machine in my home, while may be optimal from a cost minimization perspective, compromise a bit too much on quality-of-life for me—not simply from a developer perspective, but I also live somewhat of a nomadic lifestyle and prefer to have few physical possessions.<p>I already tried applying for AWS Activate but was rejected due to not meeting their criteria. I do not have a company or non-profit entity formally set up for this project. It initially started as a research project purely for personal interest, to explore how LLMs can be used for memory augmentation (in general, not just for language learning). I’ve been a user of conventional SRS tools such as Anki for many years, and have been frustrated by certain UX limitations that hindered consistent long-term use. Initially, I built it as a tool for myself, with me as the sole target audience. Onboarding foreign monks in Thailand was a fortuitous occurrence that I had not initially anticipated.
iamflimflam14 个月前
Without any information on your stack it’s very hard for anyone to give advice.<p>One thing that might be worth looking at is moving your backend into a lamda. There are various shims that make this very easy for node based backends.<p>For low traffic websites (even for quite high traffic sites) this can be almost free.<p>Your main costs will then be whatever database you are using.
amluto4 个月前
This probably doesn’t make sense at your particular scale, but colo space can be surprisingly inexpensive, and it’s kind of fun at least until you get bored. In my local not-particularly-cheap market, a full rack with 120V&#x2F;20A, 1Gbps Internet, and more cooling than you probably need costs about $400&#x2F;mo. There is no particular requirement to put rack mount servers in it, although something with remote management capability is probably a good idea.<p>You can fit a <i>lot</i> of storage in there, and while you won’t get AWS-level availability or durability, the price, latency and IOPS will blow AWS out if the water, as long as you&#x27;re using those IOPS and bandwidth within the same facility. Similarly, you can fit quite a bit of compute, but you’ll start needing to pay for more power if you need more than a couple servers worth.<p>At your scale, Hetzner is probably a better idea.
rvz4 个月前
&gt; Right now I&#x27;m running everything on AWS, but the cost feels overkill for my current needs. Since my past professional background is in AWS, it was the default choice, but given that this project is currently funded through my own savings and I&#x27;m currently unemployed, I&#x27;m looking to minimize costs. Scalability isn&#x27;t a concern as the user base is small.<p>Given that this doesn&#x27;t make any money, you might as well use local or self-hosted LLMs instead of cloud based LLMs.<p>AWS for this use-case (especially when you are not intending to make money) is indeed overkill.
randometc4 个月前
What AWS services are you using currently and what drives your costs?<p>We’re a nonprofit with apps that see ~1000s of users a month, but mostly in the US. I think we see good savings using Google Cloud Run and scaling down when there’s less traffic. You could probably set up AWS Fargate similarly. Modern app frameworks start quickly so cold starts aren’t terrible. Docker containers are portable if you outgrow that kind of environment and want to shift to dedicated VMs or the other way to serverless in future. I would also look at fly.io.
krab4 个月前
It depends on your needs for reliability and how many users you have.<p>If you don&#x27;t mind being offline when doing maintenance occasionally and being down on the rare occasion when the hardware fails (like once in a few years) until you get the replacement part, then look up &quot;server housing&quot;. You can buy a computer, either rack mounted, or tower, maybe used, and you put it into a datacenter that manages bandwidth, power and cooling.<p>That would be probably the cheapest option but pick something close to where you live.
lsb4 个月前
I’ve been running half-a-billion parameter models comfortably in a web browser, especially with WebGPU, and you can definitely run billion parameter LLMs in the browser. It becomes a heavyweight browser app, but if the main costs are running ML models you can pretty easily serve static files from a directory and let clients’ browsers do the heavy lifting. Feel free to reach out if you have questions, happy to help, I’ve been working on language web apps as well
dcminter4 个月前
I&#x27;d think in terms of running this on a PC in my spare room if I was trying to keep costs down and the usage wasn&#x27;t going to be all that high. Depending on how many concurrent users you&#x27;re expecting, and if you do something with open source models, then Ollama running on a very not-top-tier GPU is likely to be a good bet.
gumperbumper4 个月前
Can’t you run one of AWS’ t-series servers for like $10 a month? If you’re just making a few API calls, then that should be more than sufficient. My advice: don’t prematurely optimize. You _may want_ to explore self-serving models. But you also may not. Wait till you have a real problem before trying to solve it.
lloydjones4 个月前
Port it to using SST (<a href="https:&#x2F;&#x2F;sst.dev">https:&#x2F;&#x2F;sst.dev</a>) then you can more easily experiment with different cloud providers’ costs by deploying to A, B and C service.<p>SST uses Pulumi so (per other comments) there are constructs for Hetzner among other budget hosts.
评论 #42823102 未加载
rekl4 个月前
Fastly has the Fast Forward program. Seeing as your app is non-profit, perhaps their program could cover part of your infrastructure needs?<p><a href="https:&#x2F;&#x2F;www.fastly.com&#x2F;fast-forward" rel="nofollow">https:&#x2F;&#x2F;www.fastly.com&#x2F;fast-forward</a>
moltar4 个月前
Stay on AWS and apply for $1K credits. Just position yourself as a potential money making startup on the Activate application. It’s very easy to get.<p>Then you can get $5K ($4K actual) package by using a code from Mercury bank or Secret plan purchase ($200).
huksley4 个月前
I also planning to move workloads from public clouds - it is expensive and too complicated.<p>I created a service to easily deploy modern apps to VPS. Like Vercel but separating control plane and servers.<p>Supporting both CPU and GPU servers.<p>Similar to Coolify but much easier and cheaper to use.
评论 #42824290 未加载
评论 #42823794 未加载
qbqetell4 个月前
We don&#x27;t have a big brand name yet, but would you be open to using our platform? I&#x27;m the founder and you would have direct access to me at all times.
Maro4 个月前
Sign up for a $25&#x2F;mo dedicated server on OVHCloud, you can use it as a cloud devbox, host your blog, ssh-tunnel for VPN, run projects like this on it, etc. You can still call AWS APIs if you want.
zachlatta4 个月前
We are a nonprofit and are in the process of switching all of our hosting over to Coolify. It&#x27;s been great. We&#x27;ve moved 13 apps over to it so far.
pryelluw4 个月前
How many users and bandwidth do you estimate to have and use ? Do you store a lot of assets (images, video, sound)?
pottash4 个月前
Open-source models sounds like it would be cheapest, maybe running it on a VPS or something
morphle4 个月前
$3.19 or 3,04 Euro excluding VAT per month for all website server, data and labour costs (1-7) over several years, including running the big LLM (as tested, with some large LLM&#x27;s and MLX&#x2F;EXO you might need dual 64GB servers[2], doubling the monthly cost).<p>Explanation of the details:<p>For 33 years[1] I&#x27;ve hosted several huge websites from several servers in several peoples homes for less than 5 euro per month. As a professional datacenter builder and as ISP I&#x27;ve never found a cheaper option and that is not an opinion but a deeply and constantly researched measurement based on hundreds of servers. In 2025 we lowered the price to 3 Euro per month. By sharing a server with other customers, we can lower the cost to below 1 euro per month (redundant backup servers double this cost).<p>In short, you have these costs:<p>1) Energy cost per kWh per year from a supplier or the levelized cost of electricity (LCOE) of a solar&#x2F;battery, windmill&#x2F;battery or hydro&#x2F;battery system.<p>2) ISP or datacenter cost per Mbit&#x2F;s per year for the internet connection, fiber optic or copper leased line, satellite.<p>3) Colocation or closet&#x2F;attic&#x2F;barn space cost per year (with redundant servers you need several spaces)..<p>4) Server hardware cost per year based on the Total Cost of Ownership (over the lifetime).<p>5) Labour setting up and maintaining the software on the server (excluding content) per year.<p>6) Labour setting up and maintaining the content (excluding software and maintenance cost) per year.<p>7) Domain name registration, IP number registration, certificate cost.<p>8) Cooling cost[3].<p>The energy cost is highest of the world in my region (EU, The Netherlands) because of the inflating energy prices per kWh since November 2021. So you move the server to another home or datacenter with the lowest kWh price of the region (7.0 Euro cent from a hydro dam in rural Spain). Or you install solar panels and batteries (1.2 Euro cent per kWh LCOE).<p>With an €479 M4 Mac mini server in 2024 I lowered my server&#x27;s power from an average 12 watt (a 2006 Mac mini using 10.4-30.6 watt and in period 1993-2006 a 20.2-31.3 watt server) to today (januari 2025) a 4.0-4.7 watt with an expected lifetime of 12-14 years. The server power includes the modem and router power. If you need a few terabytes disk space your cost per month will go up by around a euro per month.<p>In 2025 I expect the customer wil pay €37.21 for a year redundant availability of 10 TB at 1 Gbps for IP transit traffic per month, 16 GB DRAM unix server with 10 cores, 10 GPU cores and 16 neural engine (unlimited websites, email). If you would max out the server&#x27;s 43 Trillion operations per second at max performance the power cost might go up 10x. Almost 3 euro per month for a server that can run an LLM continuously.....<p>You might hear of hosted servers below 3 euro. They usually involve some discount scheme to lure new customers in. With a minimum price of €7 for the cheapest domain name and certificate registration (yearly, worldwide), servers maintenance labour of €30 per hour these price quotes below 3 euro per month usually are not based on true cost.<p>[1] In the perod 1986-1992 we hosted email, FTP and Gopher on the university&#x27;s unix servers with UUCP and TCP&#x2F;IP. Since 1992 I hosted on the TCP&#x2F;IP servers in my home, my friends home and several offices. Since 1999 we host in the EU, US, Canada with options for Brazil and Asia.<p>[2] M4 Mac Mini Cluster <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=GBR6pHZ68Ho" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=GBR6pHZ68Ho</a><p>Anyone with experience or proof of a cheaper solution? Please contact me so we can bid on your knowledge.<p>[3] I heat my rooms with several servers, so the cooling and fan cost are $0, even in summer during heatwaves.
评论 #42823628 未加载
评论 #42919981 未加载
moltar4 个月前
Maybe explain your stack and why it’s expensive I can help suggest saving strategies.