Alexis Bouchez

~ / research / Startup Ideas Hidden in the Railway Station Forum

Startup Ideas Hidden in the Railway Station Forum

Railway has 2.8M developers and 72,796 threads on its public forum at station.railway.com. I pulled the sitemap (1,354 feedback URLs), picked the high-signal ones, and extracted verbatim quotes so you can read what users actually say.

Most of these ideas are either:

  1. Things Railway said publicly they will not build (and recommends a third-party for).
  2. Things Railway has been promising for 2+ years with no ETA.
  3. Things that could exist as a portable tool across Render, Fly, Koyeb, and Railway, not just Railway.

Each idea below links to the actual thread so you can read the context before you build.

1. How I scraped it

  • Fetched https://station.railway.com/sitemap.xml, which splits into three 60k-entry sub-sitemaps.
  • Parsed all URLs matching /feedback/ pattern: 1,354 feedback threads indexed.
  • Filtered slugs by keywords (billing, backup, log, metric, region, etc.) and fetched the highest-signal ones.
  • Quotes below are pulled verbatim from the public threads.

2. The 15 startup ideas

1. Cost Explorer for Railway

Users want an AWS Cost Explorer–style view of Railway spend, per-project and across billing cycles. Railway's native billing page is not detailed enough for teams, and the in-dashboard cost calculator was quietly removed.

Quotes:

  • "I'd like to go back and compare the end-of-month amounts per-project." (shixzie)
  • "Something like the AWS Cost Explorer dashboard (no need to be that fancy) would be nice." (shixzie)
  • "The ability to view billing and usage for past cycles." (brody, Railway employee, agreeing)
  • "it would be nice to have some calculator on the dashboard that could estimate how much it would cost to host on railway" (estimated bill thread)
  • "They removed this?" / "I can't see it on the page at all but was sure I had seen it before" (on the now-removed calculator)

What to build: read-only dashboard that reads the Railway GraphQL API, breaks down cost by project / service / environment, shows MoM deltas, flags anomalies, and lets finance people export CSVs. Bonus: forecast next bill from the current month's trajectory. Charge $29/mo per workspace. Adjacent markets: same product can plug into Render and Fly by the end of month 2.

Evidence of demand: thread is Open, moderators have chimed in acknowledging the gap but not committing.

2. Railway Backup Vault

Database backups are a Pro feature but users cannot download them. They also cannot label them. They also cannot schedule them from the CLI in the way they want. This is a full mini-product sitting inside a feature request.

Quotes:

  • "it seems kind of useless if I am not allowed to download the backup for locally" (download backups thread)
  • "a pg_dump export would be nice too"
  • "Allow labeling the backup with that reason, so that future viewers can understand why a manual backup was created" (label backups thread)

What to build: a scheduled backup service that runs pg_dump / mysqldump against Railway-hosted DBs, stores compressed/encrypted backups in the customer's S3 or R2, supports labels, retention policies, and one-click restore. Think "TablePlus + dbBackups.io, but for Railway (and Render, and Fly)". Charge per-GB stored plus a flat monthly.

Evidence: status is "under review". Railway will not ship this fast; they are infra first, DX second on backups.

3. Status Page / Incident Communication (explicit Railway punt)

Railway publicly said they will not build this.

Quotes:

What to build: a status page product with a first-class Railway integration that auto-probes health checks, listens for deployment events via webhooks, and correlates with Railway's own outage feed so customers do not have to manually write "we are investigating" when Railway itself is down. Instatus already exists but the tight Railway integration is an open lane. Price: $19/mo starter.

4. AI Log Search for Railway

Users want to ask natural-language questions against their logs without switching tools.

Quotes:

  • "I should be able to ask questions about something going on in my system to the logs so it can find any relevant information for me." (AI log search thread)
  • "logging is really fucking broken / Sometimes logs are cutoff / fail to show" (better logging thread)
  • "I would love to start migrating our apps from Heroku to Railway, but Logtail support would be essential"

What to build: lightweight log ingester that hooks into Railway via log drains (now supported), stores 30 days of logs in ClickHouse, and lets users query in plain English ("when did the checkout service 500 last week?"). Add PagerDuty/Slack alerts based on saved queries. This is Sentry + Logtail + Bolt.new mashed into one. Can be built in 2 weeks on top of OpenAI's embeddings + a Postgres vector column.

5. Deployment Alerting Bridge

Railway does not emit deployment-completed webhooks in a way that teams with complex rollouts can rely on.

Quotes:

  • "the rollout strategy they may have is a bit complex that may have multiple steps such as migrations" (deploy alerting thread)
  • "workspaces managing multiple services may lack visibility into overall environment health during deployments"
  • "difficult to communicate deployment completion to their team"

What to build: polling + webhook middleware that subscribes to Railway's API, waits until every service in an environment reports ready/failed, then fires a single "environment healthy" notification into Slack, Teams, Linear, or a custom webhook. Sell as "DeployStatus for Railway", $15/mo per project. You can generalize to any PaaS.

6. Railway Observability Drain (Prometheus + OTel)

Users want to scrape metrics off Railway DBs and services into their own Prom/Grafana stack.

Quotes:

  • "Additional Option to Enable Metrics Endpoint or allow developers for opt in their own Metrics solution to use in Prometheus" (metrics endpoint thread)
  • "Custom metrics in a project which have got Railway's beautiful metric viewer." (custom metrics API thread)
  • "This would be sent via an API, or an SDK if that gets created."
  • "Bumping this, would be awesome." (kajdev)

What to build: sidecar template plus SaaS that (a) runs an exporter pod against Railway Postgres/Redis/MySQL, (b) pushes into a managed OTel/Prometheus endpoint, (c) bundles dashboards and alert rules. Open-source the sidecar template; charge for the managed collector. Grafana Cloud has the infra but not the Railway ergonomics.

7. RBAC / Environment Access Control Service

Railway has confirmed they do not have RBAC on the roadmap.

Quotes:

  • "Access control is at the project level, meaning anyone with access can modify Production services" (env access thread)
  • "We need more granular access control to restrict developers from making changes to Production"
  • "We don't currently have any projects on the board for RBAC so unfortunately, I cannot offer an ETA" (Railway employee)

What to build: a wrapper layer that proxies Railway API calls, enforces an RBAC policy (who can deploy to prod, who can edit env vars, who can read logs), and logs every action to an audit trail. Ship a CLI replacement and a GitHub Action. Sell to startups that need SOC2 controls but outgrew seat-level permissions. $49/mo seat, audit log included.

8. Project Snapshot / IaC Export

Railway projects cannot be exported today. Users want a .railway file they can commit.

Quotes:

  • "Create a project snapshot that would restore all our project settings, services, parameters, instances, variables, (everything)" (project snapshots thread)
  • "The ability to export an .railway file that is only compatible with Railway to save a copy of our project settings"
  • "I would be more than happy to make a couple of snapshots every few days and pay the GB/min of the snapshot files"

What to build: railway-export CLI that inspects a Railway project via the API and writes a Terraform module or a Pulumi program. Bonus: weekly automated snapshots to the user's Git repo. Charge $9/mo per workspace or open-source the CLI and SaaS the automation. Gives teams disaster-recovery and PR-reviewable infra.

9. Railway Volume Manager

Railway volumes exist but you cannot browse, upload, or download the files inside them from the UI.

Quotes:

  • "Please, include the ability to upload (and download). I want to import some volumes from another provider" (ryancasas, volume browser thread)
  • "upload and downloading files using the CLI would feel nice... the ability to 'mount' the volume into a local folder" (devparapalli)
  • "I had to write custom code to manage sqlite db files, I'd love to scrap this" (dkossman)
  • "Such a feature would change my life" (dev, moderator)

What to build: a file browser UI that mounts over Railway volumes via the GraphQL API, lets users upload, download, preview, and rsync. Think "Cyberduck for Railway volumes". $10/mo. SQLite users will pay on day one.

10. PR Environment Suite with Custom Domains

PR environments on Railway are under-loved and missing custom-domain support.

Quotes:

  • "We know you all have thoughts about how PR environments should work (and environments in general), and we would love if you would take a moment to let us know." (Railway team, PR environments thread)
  • Users want https://myapp-pr-123.mydomain.com instead of https://myapp-pr-123.up.railway.app (custom PR domains thread)
  • Named use cases: "Eliminating cross-origin concerns" and "Enabling shared cookies across preview environments"

What to build: middleware that spins up per-PR Railway environments off your GitHub branch, provisions custom subdomains with wildcard TLS on your domain, injects preview-specific env vars, and tears down on PR close. Bundle a "preview comments" bot that posts to the PR with a live URL + Lighthouse scores. Vercel has this for Next.js apps; you do it for arbitrary Dockerfiles + Railway. $25/mo per GitHub repo.

11. Cron / Scheduled Jobs Service

Users keep asking for pg_cron and other scheduling extensions.

Quotes:

  • "I really don't want to create a separate image from your official one just for this extension." (pg_cron thread)

What to build: a Railway-native cron runner that accepts HTTP targets, shell commands, or SQL queries, schedules them via standard cron syntax, retries on failure, and emits metrics. Competes with Upstash QStash and Trigger.dev, but cheaper and pre-configured for Railway users. $7/mo for 100 scheduled jobs.

12. Template Marketplace with Locked Variables

Railway templates are huge (there is a new-templates section on the home page), but creators cannot lock critical env vars, and there is no paid-template distribution.

Quotes:

  • "It would be great to make this variable hidden/locked when a user tries to deploy with my starter button." (locked env thread)

What to build: a commercial template marketplace where creators publish Railway starters, protect secrets, and charge per deploy (say $5 per install). Railway already said yes to a $ symbol next to templates; building the marketplace lets indie hackers monetize stacks. Take 20%. See it as "Gumroad for infra starters".

13. Approval Gates / GitOps for Deploys

Railway auto-deploys are on or off, with no between.

Quotes:

  • "We'd like to deploy staging env automatically on merge but having a manual approve of production env" (manual trigger thread)
  • "When automatic deploys are disabled - how can we deploy manually through UI?" (grundmanise)

What to build: an approval layer that sits between GitHub merges and Railway deployments. When a PR merges to main, your product waits for an approver click in Slack, then triggers the Railway deploy via API. Perfect for teams transitioning from Heroku pipelines. $19/mo per team.

14. Build Cache as a Service

Railway users notice slow deploys and want build-step caching.

Quotes:

  • "it would be great to have some sort of build caching of dependencies to speed up deploys" (build cache thread)
  • Context: Rails + Nixpacks users are the loudest; a single-digit-minutes improvement would move real money for them.

What to build: a remote caching plane for language-specific artifacts (Ruby bundles, node_modules, Go modules, Python wheels) that Railway builds pull from before full install. Integrates via a 5-line Dockerfile shim or a Railpack plugin. Sell to Railway, then generalize to Fly and Render. Stripe pricing: $0.10 per cached-build GB month.

15. Secure CLI / Audit Gateway (2FA + signed commands)

Basic 2FA exists but there is no audit log or signed-command flow for destructive operations.

Quotes:

  • "Providing accounts with 2FA will increase security on accounts. This is great considering the types of sensitive information that can be exposed in dashboards." (2FA thread, completed)
  • "Would be useful for sensitive actions like wiping plugin data/deleting projects."

What to build: a Railway-aware CLI proxy that requires WebAuthn signatures for production deploys, project deletes, or env var changes. Emit a tamper-evident audit log. Pitch: "SOC2 Type II control for your Railway account in 10 minutes." $99/mo per workspace, compliance add-on.

3. Three more bonus ideas

16. Multi-region Smart Router (for teams Railway cannot serve yet)

Every region thread is essentially a ticket that says "we cannot use Railway because latency to our users is too high." Rather than wait for Railway to open São Paulo, Africa, Australia, and India, build a Cloudflare-Workers-powered router that fronts a user's Railway deployments with edge-cached responses and session affinity. Sell it as "Railway Global". Threads: Configurable Deployment Region, India region, Australia, Africa, Brazil. Quote: "Can't use Railway without a region in europe, due to GDPR and other concerns."

17. Railway GPU Broker

Railway Metal threads show that GPU is the #1 blocker for AI teams.

Quote: "GPU support is a MUST have for me. Railway is great for prototyping, but I need GPU otherwise I'm moving to fly.io." (Railway Metal feedback)

Build: a broker that provisions GPU capacity on RunPod, Lambda, or CoreWeave, fronts it with a Railway-style deploy UX, and lets users keep their app on Railway while offloading inference to GPUs elsewhere. Per-hour markup or flat SaaS.

18. Affiliate / Referral Tracking Platform for PaaS

Railway's own affiliate program is a "Completed" thread; the referral mechanic works but there is no third-party dashboard for people who push high volume.

Quote: "You get 15% commission on every dollar your referrals spend on Railway within the first 12 months." (affiliate thread)

Build: a tracking dashboard for creators/agencies with multiple referral links across Railway, Vercel, Render, Fly, Supabase, and Neon. Pulls payout data, attributes clicks, predicts MRR. The Rewardful / PartnerStack of cloud-referral money. Charge 2% of payouts.

4. Patterns I noticed across 1,354 threads

  • Billing is the single biggest open pain cluster. Threads: bill calculator (completed but removed), billing admin, billing-per-project, change-billing-date, better-billing-reporting. Finance ops is underbuilt on Railway.
  • "Just integrate with X" is the recurring resolution pattern. Railway politely declines to build status pages, dedicated monitoring, or 3rd-party log sinks and points at other tools. Every decline is a distribution moat for a focused startup.
  • Database ergonomics are under-served. Backup labels, downloads, restore, DB column UI, cron, metrics. Railway ships infra; tooling around it is a long tail.
  • Enterprise concerns (RBAC, SSO, audit, regions) have no ETA per Railway staff replies. If you sell security-layer SaaS, every Railway customer over 50 seats is a lead.
  • PR environments are explicitly asking for community input. A bolt-on product can ride the coattails of whatever Railway ships.
  • Templates are Railway's distribution engine (showcased on the home page) but monetization is missing; a paid-template marketplace has a captive audience.

5. How to pick one

Rank each idea against these filters:

  1. Did Railway explicitly say "no" or "we recommend X"? ✅ You get a free pass. (Status pages, Instatus model.)
  2. Has the thread been "Under Review" for 12+ months? ✅ Railway is not prioritizing it. (Backup labels, custom metrics API, RBAC, PR env domains.)
  3. Is the request solvable with a thin wrapper on the public Railway API? ✅ Weekend project, ship in 2 weeks. (Cost explorer, export, volume browser, approval gates.)
  4. Does the same product sell on Render / Fly / Koyeb? ✅ Your TAM is not capped by Railway's size. (Backups, logs, status, cron, RBAC, build cache.)

The three I would pick if starting tomorrow:

  1. Cost Explorer for Railway (low effort, high universal pain, easy to port).
  2. Railway Backup Vault (Railway is not fast here, security-adjacent, easy upsell).
  3. Railway AI Log Search (the AI-wave tailwind + the ergonomic gap Railway will not fill).

6. Data appendix

  • Sitemap fetched 2026-04-13 from https://station.railway.com/sitemap.xml.
  • 1,354 /feedback/ URLs, 3 shards of 60k lines each.
  • Platform stats from homepage: 72,796 threads, 2,823,271 registered developers.
  • Every quote above is verbatim from the linked thread; click through to read the full context before building.

Update cadence: rescrape quarterly. A lot of these threads will move from "Under Review" to "Completed" as Railway's roadmap lands, and when they do, the startup idea evaporates. Speed matters.