Deploy and Host Cap with Railway

Cap is an open-source screen recorder: record your screen and camera, get a shareable link right away, and collect comments on the video. This community template runs Cap's web app and ffmpeg media server on Railway with MySQL and a Railway Bucket for the recordings, so your videos stay in your own project.

About Hosting Cap

Cap is a Next.js web app that hosts the dashboard, share pages and the API used by Cap Desktop, plus a media server that runs ffmpeg for transcoding, thumbnails and previews. It needs MySQL for accounts and video metadata and S3-compatible storage for the files, which browsers and the desktop app upload to directly. This template uses Cap's official images pinned by digest, a Railway Bucket with the CORS rule browser uploads need, a private media server with a generated callback secret, a database wait so first boot does not race MySQL, and a small cron job that retries failed video processing. All secrets are generated at deploy time.

Common Use Cases

Async updates and stand-ups recorded in a few minutes instead of long written reports
Bug reports that show the problem on screen, linked from your issue tracker
Onboarding, support and how-to walkthroughs on a URL you control
Recordings of customer or internal systems that must not leave your own infrastructure

Dependencies for Cap Hosting

Cap web app (ghcr.io/capsoftware/cap-web, digest-pinned)
Cap media server with ffmpeg (ghcr.io/capsoftware/cap-media-server, digest-pinned)
MySQL (Railway MySQL 9.7)
S3-compatible object storage (Railway Bucket)
Optional: a Resend account for login and notification e-mails

Deployment Dependencies

Cap self-hosting guide: https://cap.so/docs/self-hosting
Cap source and Docker Compose files: https://github.com/CapSoftware/Cap
Environment variable reference: https://github.com/CapSoftware/Cap/blob/main/packages/env/server.ts
Resend (e-mail API): https://resend.com
Railway Buckets: https://docs.railway.com/storage-buckets

Implementation Details

Service	Image	Role
Cap	`ghcr.io/capsoftware/cap-web` (digest)	Web app and API on the public domain; runs database migrations at boot
Cap Media Server	`ghcr.io/capsoftware/cap-media-server` (digest)	ffmpeg processing, private network only
MySQL	`mysql:9.7.2`	Accounts, videos, comments (volume, no public proxy)
Bucket	Railway Bucket	Recordings, thumbnails, exports
Bucket CORS	`amazon/aws-cli:2.37.12`	One-shot job that sets the bucket CORS rule for Cap's domain
Cap Maintenance	`curlimages/curl:8.22.0`	Cron every 15 minutes: retries failed or stalled video processing

First login: open the public URL and enter your e-mail. Until you add a Resend API key, the 6-digit login code is written to the Cap service logs; copy it from there. Before inviting others, set RESEND_API_KEY and RESEND_FROM_DOMAIN (Resend is an HTTP API, so it works on every Railway plan) and optionally restrict sign-ups with CAP_ALLOWED_SIGNUP_DOMAINS. In Cap Desktop, set the server URL to https://{your-domain}.

Custom domain: add it to the Cap service, update WEB_URL, add https://{your-domain} to CORS_CONFIGURATION on Bucket CORS, and redeploy both.

Scaling: give Cap Media Server more CPU and memory if many recordings are processed at once; it adapts its concurrency to the resources it sees. Keep Cap at one replica (it migrates at boot) and scale it vertically.

Versions: Cap publishes rolling latest images, so this template pins image digests. To upgrade, set new digests on Cap and Cap Media Server together, back up MySQL, and redeploy; migrations run on boot.

Why Deploy Cap on Railway?

Railway runs the web app, media server, database and bucket as one project with private networking, managed volumes and usage-based billing. Video playback and uploads go straight to the bucket, so you get a self-hosted Loom alternative without managing servers, certificates or storage yourself.