How Makers Build
Check my stackPlan my stack
Menu

Amazon Web Services vs Railway

Two sides of the hosting decision: hyperscaler cloud and paaS. When each fits, what it costs, who moves from one to the other, and what makers who chose it say.

Amazon Web ServicesHyperscaler cloud
RailwayPaaSllms.txtMCP
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

Which fits you

Choose Amazon Web Services if
  • The widest catalog of building blocks — Lambda, containers, managed databases, queues — to assemble your own infrastructure.

Use it whenYou need a specific AWS service, enterprise customers ask for it, or you have credits to spend.

Trade-offYou wire up networking, permissions and deploys yourself; billing is per service and hard to predict.

Choose Railway if
  • You have a long-running server, a background worker and a database to deploy together

Use it whenYour app is a server plus a database and a worker, and you want them in one project.

Trade-offUsage-based billing needs watching; fewer regions and compliance options than a hyperscaler.

At a glance

Amazon Web ServicesRailway
Used by386 makers' products · 17 open-source projects112 makers' products
Cost at default usagerequests 5 million requests, data transfer out 100 GB, serverless function compute 100 GB-hours$17/mo Pay as you go (Lambda + EC2 t4g.small)$45/mo Hobby
Moved to it on GitHubpull requests since Oct 20245 from Railway9 from Amazon Web Services
Downloads4.2M/wk−39% vs npm
PricingPay-as-you-go per service; new accounts get up to $200 in Free Tier credits over 6 months. · paid from Pay as you goUsage-based with a small monthly minimum. · paid from $5/mo minimum usage
Free tierYesNo
Open sourceNoNo

Cost as you grow

At 100k requests Amazon Web Services costs less ($16 vs $40); from about 50M requests Railway does ($90 vs $160). They're different kinds of tool — hyperscaler cloud and paaS — so the prices don't buy the same thing.

$0$50$200$500$1,0000.10.5151050100500
RailwayAmazon Web Servicesx: requests per month (million requests), other usage scaled with it; held at 1 server (always-on app servers) · cheapest usable plan at each point, list prices · try your own numbers
The numbers, plan by plan
Requests per monthAmazon Web ServicesRailway
0.1$16 Pay as you go (Lambda + EC2 t4g.small)$40 Hobby
0.5$16 Pay as you go (Lambda + EC2 t4g.small)$41 Hobby
1$16 Pay as you go (Lambda + EC2 t4g.small)$41 Hobby
5$17 Pay as you go (Lambda + EC2 t4g.small)$45 Hobby
10$32 Pay as you go (Lambda + EC2 t4g.small)$50 Hobby
50$160 Pay as you go (Lambda + EC2 t4g.small)$90 Hobby
100$320 Pay as you go (Lambda + EC2 t4g.small)$140 Hobby
500$1,600 Pay as you go (Lambda + EC2 t4g.small)$540 Hobby

From each vendor's pricing page: Amazon Web Services, Railway.

Who moves from one to the other

Public pull requests on GitHub since Oct 2024 whose title says "Amazon Web Services to Railway" or the reverse — real code changes, by developers in general rather than makers only.

Railway → Amazon Web Services5 PRs
All matching pull requests on GitHub ↗
Amazon Web Services → Railway9 PRs
All matching pull requests on GitHub ↗

What makers say

Makers on using it for hosting, from Product Hunt and Starter Story interviews, each linked to the source. Products with a page of their own and fuller notes first.

On Amazon Web Services
We are so excited to integrate with AWS Lambda. Look for our upcoming co-authored blog about how Fleak and AWS Lambda bring goodness to data teams and help them win their time back!
Fleak, the makerSep 2026 ↗
AWS provides the global infrastructure we need to deliver low-latency voice translation to users worldwide. Their reliable services ensure our calls connect seamlessly across regions.
The Banana App, the makerSep 2026 ↗
AWS Lambda's scalability is definitely a game-changer, especially for handling the traffic spikes from a Product Hunt launch. Shoutout to AWS Lambda for making it possible! 🚀
Gan.AI, the makerSep 2026 ↗
152 more on the Amazon Web Services page →
On Railway
Railway makes deploying and scaling our backend effortless. Push to main, it's live. No Dockerfiles to debug, no infra to babysit. We went from zero to production in an afternoon.
Agentcard, the makerSep 2026 ↗
🚀 Shoutout to Railway for making API deployment a breeze! As a frontend developer without much backend experience, their platform made it super easy to set up and deploy changes.
npmpackage.info, the makerSep 2026 ↗
Railway is simple and handles deployment transitions really well. Old deployments drain cleanly while new ones come up, so users are never interrupted when we ship.
Sup AI, the makerMar 2026 ↗
60 more on the Railway page →

Loved and watch-outs

Themes that recur in makers' words and Hacker News comments, each linked to what it summarises.

Amazon Web Services
Most loved
  • It offers a huge catalog of services in one place, from databases to queues to AI. PHPH 2PH 3PH 4
  • It scales up and down with bursty traffic without re-architecting. PHPH 2PH 3
  • Its reliability record is something teams are willing to bet a product on. PHPH 2PH 3PH 4
Watch-outs
  • Major outages, from long us-east incidents to struck data centers, have taken customer services down. HNHN 2HN 3HN 4
  • Its services can be tricky to configure. PH
On Product Hunt: 5.0★, 282 reviews · mentioned most: reliable cloud services, scalable infrastructure, extensive service offerings
Railway
Most loved
  • A git push deploys a backend to production in minutes with no DevOps work. PHPH 2PH 3PH 4
  • Services, Postgres databases and buckets live in one place, with a marketplace of ready integrations. PHPH 2PH 3PH 4
  • Pay-per-resource pricing keeps small and self-hosted services cheap. PHPH 2PH 3HN
Watch-outs
  • Platform-wide outages, including one when Google Cloud blocked Railway's account, took customer services down. HNHN 2HN 3HN 4
  • It adds a middle layer on Google Cloud that controls your availability, features and security. HNHN 2
  • API tokens grant every operation without clear warnings, which let an agent instantly delete a production database. HNHN 2
On Product Hunt: 5.0★, 103 reviews · mentioned most: fast deployment, scalable infrastructure, no deep DevOps knowledge required

Who uses each

Amazon Web Services386 makers' products
Ada.imAgentMailAgentplace.ioAgentRAI Context FlowAlchemyzeAmplooAntiquotedAnythingAstridAstrioBeatoven.ai+374 more
Railway112 makers' products
24CallDeskAgentcardAgentRailAmamiAprilAviquillAximote In-Car AppBlink.newBreadcrumbChat ThingClawSecureClick+100 more
Used by both — often one replacing the other, or each for a different part of the product
BetterClawNotteSymvol Video AI for EducationBlink.newReloopMarked Share

What makers pair each with

With Amazon Web Services
Transactional Email
LLM API
Frontend Framework
Payments
With Railway
Frontend Framework
Database
Transactional Email
LLM API

Neither quite right?

See the whole hosting decision →
VercelFrontend cloudNext.js apps — zero-config deploys and preview URLs.vs Amazon Web Services →vs Railway →
NetlifyFrontend cloudStatic sites and JAMstack apps on any framework.vs Railway →
CloudflareEdge platformCheap global hosting with Workers, plus database (D1) and storage (R2) on the same platform.vs Amazon Web Services →vs Railway →
RenderPaaSA Heroku-style platform for web services, cron jobs and Postgres.vs Amazon Web Services →vs Railway →
Fly.ioContainer platformDocker apps that should run close to users in many regions.vs Amazon Web Services →vs Railway →
Hetzner CloudYour own serverA lot of CPU and RAM per euro, in European, US and Singapore regions.vs Amazon Web Services →vs Railway →
CoolifyYour own serverA Heroku- or Vercel-style dashboard on your own server, with many one-click services.vs Railway →