Deploy and Host Nginx on Railway

Built for an LLM agent to deploy and use. There is no content to bake into an image and no form to fill in: the site lives on a volume, an upload token is generated at deploy time, and the contract below lists every URL, variable name and the one script that proves it works. Publish a site with curl, no git repository and no redeploy.

Nginx is the open-source (BSD-2) web server and reverse proxy behind a large share of the web. This template runs the official Nginx 1.30.6 image (stable branch, pinned by digest) as a static site host: files under /data/www on the volume are served on the public domain, and the built-in WebDAV module accepts PUT, DELETE, MKCOL, COPY and MOVE from anyone who presents the upload token.

About Hosting Nginx

This template runs one service with a volume at /data. On every boot the start command writes the nginx server block and the upload credential from the service variables, so the config never drifts from them. /data/www is the web root (a placeholder index.html is written only when none exists), /data/tmp holds upload temp files on the same volume so a PUT can be renamed into place, and /data/nginx.d is an extension point: any *.conf file there is included inside the server block on every boot, which is where location blocks with proxy_pass to other services in the project go. Reads (GET, HEAD) are public; every write needs HTTP basic auth with the user deploy and the password UPLOAD_TOKEN. Rotating the variable rotates the credential on the next deploy.

Agent contract

Public URL: https:// plus the RAILWAY_PUBLIC_DOMAIN of the nginx service, also published as the variable SITE_URL
Private URL: http://nginx.railway.internal:8080 (also the variable SITE_PRIVATE_URL), for other services in the project
Credentials: UPLOAD_TOKEN (user deploy). It stays the same across redeploys.
Auth: HTTP basic auth, -u deploy:$UPLOAD_TOKEN, on every method except GET and HEAD. A write with no or a wrong credential returns 401.
Health: GET /healthz returns 200 with the body ok, no credentials

# run with SITE_URL and UPLOAD_TOKEN set
curl -s -o /dev/null -w 'health %{http_code}\n' "$SITE_URL/healthz"
curl -s -o /dev/null -w 'put-no-credential %{http_code}\n' -X PUT -d x "$SITE_URL/docs/a.txt"
curl -s -o /dev/null -w 'put-wrong-token %{http_code}\n' -u deploy:wrong -X PUT --data-binary 'hello from the agent' "$SITE_URL/docs/a.txt"
curl -s -o /dev/null -w 'put-with-token %{http_code}\n' -u "deploy:$UPLOAD_TOKEN" -X PUT --data-binary 'hello from the agent' "$SITE_URL/docs/a.html"
curl -s -o /dev/null -w 'put-index %{http_code}\n' -u "deploy:$UPLOAD_TOKEN" -X PUT --data-binary 'hello from the agent' "$SITE_URL/index.html"
curl -s "$SITE_URL/docs/a.html"
curl -s "$SITE_URL/" | grep -o 'hello from the agent'
curl -s -o /dev/null -w 'missing %{http_code}\n' "$SITE_URL/nope"
curl -s -o /dev/null -w 'delete-no-credential %{http_code}\n' -X DELETE "$SITE_URL/docs/a.html"
curl -s -o /dev/null -w 'delete-with-token %{http_code}\n' -u "deploy:$UPLOAD_TOKEN" -X DELETE "$SITE_URL/docs/a.html"
curl -s -o /dev/null -w 'after-delete %{http_code}\n' "$SITE_URL/docs/a.html"

A fresh deploy prints health 200, put-no-credential 401, put-wrong-token 401, put-with-token 201, put-index 204 (the placeholder page is replaced; 201 if you removed it first), the uploaded HTML twice (the file and the site root), missing 404, delete-no-credential 401, delete-with-token 204 and after-delete 404. Upload a whole site with a loop of curl -u deploy:$UPLOAD_TOKEN -T FILE $SITE_URL/PATH; missing directories are created.

Why Deploy Nginx on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Nginx on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.

Agent-ready - generated upload token, a contract block and one working script; no wizard and no repository to fork.
Update the site without a redeploy - upload, replace and delete files over HTTP while the service keeps running.
Files survive redeploys - the web root sits on a volume, so uploaded content is still there after a restart.
Reverse-proxy ready - drop a .conf in /data/nginx.d to proxy a path to another service over Railway private networking.
Pinned image - pinned by digest, so a redeploy never changes the server version underneath you.

Common Use Cases

Static site and landing page hosting - HTML, CSS, JavaScript and images served by nginx with gzip on.
Artifact and file hosting - an agent pushes build output, reports or datasets and shares the public URL.
Reverse proxy in front of other services - route paths to internal services from one public domain.

Dependencies for Nginx Hosting

Nothing else: the site, upload temp files and extra config all live on the volume at /data.

Deployment Dependencies

nginx/nginx - upstream project
Nginx documentation - server blocks, proxying and the WebDAV module