Deploy Medusa v2 Commerce Store (Backend + Storefront)

Deploy and Host Medusa with Railway

Launch a complete Medusa v2 store on Railway: the Medusa backend with its admin dashboard, a background worker, the official Next.js storefront, Meilisearch-powered search, Postgres, Redis and product images in a Railway Bucket. The admin user and the storefront's API key are set up automatically on the first deploy. Community template.

About Hosting Medusa

Medusa is a headless commerce platform: a Node.js backend exposes Store and Admin APIs, and any frontend can sell through it. A production setup needs more than one process. Medusa recommends running the HTTP server and a separate worker for subscribers, scheduled jobs and workflows, both backed by Postgres and Redis, with uploads in object storage. This template builds the backend and storefront from the official medusajs/dtc-starter at a pinned commit, runs migrations in a pre-deploy step, stores files in a private Railway Bucket served through a small read-only media proxy, and hands the storefront its publishable key at start-up over the private network, so nothing has to be copied by hand.

Common Use Cases

• Direct-to-consumer web shops with a custom Next.js storefront
• Headless commerce backend for mobile apps, marketplaces or B2B portals
• Multi-region stores with separate currencies, taxes and fulfilment settings
• Migration or proof of concept before moving off a hosted e-commerce platform
• Agencies building client stores on a reproducible, pinned stack

Dependencies for Medusa Hosting

• Medusa 2.21 backend (server + worker) and Next.js 15 storefront from medusajs/dtc-starter
• Postgres 18 and Redis 8.2 (Railway base services)
• Meilisearch v1.54 with @rokmohar/medusa-plugin-meilisearch
• Railway Bucket for uploads, plus a Bun media proxy for public file URLs

Deployment Dependencies

• Medusa deployment guide: https://docs.medusajs.com/learn/deployment/general
• Worker mode: https://docs.medusajs.com/learn/production/worker-mode
• S3 file provider: https://docs.medusajs.com/resources/infrastructure-modules/file/s3
• Official starter: https://github.com/medusajs/dtc-starter
• Meilisearch: https://www.meilisearch.com/docs
• Railway Buckets: https://docs.railway.com/guides/storage-buckets

Implementation Details

First login

1. Before deploying, set MEDUSA_ADMIN_EMAIL on the Backend to your address (default admin@example.com). No other input is needed.
2. Wait for the Backend pre-deploy (migrations + demo store) and for all services to turn healthy; the Storefront waits for the Backend on its first start.
3. Open https://{backend-domain}/app and log in with MEDUSA_ADMIN_EMAIL and the generated MEDUSA_ADMIN_PASSWORD from the Backend's Variables tab. Change the password.
4. Open the Storefront domain: the demo region and products are live. Add Stripe keys (STRIPE_API_KEY, STRIPE_WEBHOOK_SECRET on the Backend, NEXT_PUBLIC_STRIPE_KEY on the Storefront) when you are ready to take payments.

Scaling workers: raise the Worker's replica count in Railway; it is stateless and coordinates through Redis. The Backend and Storefront are stateless too and can run more replicas. Meilisearch, Postgres and Redis keep one instance each (they own volumes).

Search: the storefront calls Medusa's POST /store/search; Meilisearch stays private and its key never reaches the browser. To fall back to Medusa's built-in Postgres search, clear MEILISEARCH_HOST on the Backend and Worker.

Custom domains: after adding domains, update STORE_CORS, ADMIN_CORS, AUTH_CORS and MEDUSA_BACKEND_URL on the Backend, NEXT_PUBLIC_MEDUSA_BACKEND_URL and NEXT_PUBLIC_BASE_URL on the Storefront, and S3_FILE_URL if the media proxy gets its own domain.

Pinning and upgrades: the starter commit (DTC_STARTER_REF) and the Meilisearch plugin version (MEILISEARCH_PLUGIN_VERSION) are build arguments in services/backend/Dockerfile and services/storefront/Dockerfile. Change them together, check the plugin's Medusa compatibility, and redeploy; migrations run automatically. Upgrade Meilisearch with a dump/snapshot first, following its upgrade guide.

Why Deploy Medusa on Railway?

Railway runs the backend, worker, storefront and data services side by side on a private network, with pre-deploy migrations, generated secrets, volumes, a managed bucket and per-second billing. You get a production-shaped Medusa store from one deploy and can scale the API, workers and storefront independently as traffic grows.