server
{
  listen 443 ssl http2 default_server;
  server_name _;

  proxy_set_header Host $host;
  proxy_set_header Connection $http_connection;
  proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
  proxy_set_header X-Scheme $scheme;

  location ~* ^/api/(.*?|v1|openai)/(v1|(generate|stream)|(chat/completions|completions))$
  {
    # Route to inference endpoints
    proxy_pass http://127.0.0.1:5000;

    # Required for streaming (both websockets and SSE).
    proxy_buffering off;
    proxy_cache off;
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";

    # Set long timeouts for inference operations.
    # Cloudflare has a timeout of 100 seconds.
    proxy_read_timeout 120;
    proxy_connect_timeout 120;
    proxy_send_timeout 120;
  }

  location /
  {
    proxy_pass http://127.0.0.1:5000;
  }

  ssl_certificate /etc/ssl/certs/nginx-selfsigned.crt;
  ssl_certificate_key /etc/ssl/private/nginx-selfsigned.key;
  include /etc/nginx/snippets/ssl-params.conf;
}