- Launch — pick an HTTP image below (or type your own) and hit Launch. No TCP Port needed.
- Get a URL — the response includes a domain the moment the workload is ready.
- Open it — visit the URL in a browser, or
curlit.
- Launch with a TCP Port set (e.g. 6379 for Redis) — and an Environment Variable if the image needs a credential just to boot (Postgres, MySQL).
- Get a
tcpAddress— a rawhost:port, not a URL. - Connect with the service's own client —
redis-cli,psql, not a browser.
Non-HTTP support (the TCP Port field) only works on the docker engine — check the badge in the header. A published TCP address can change after a stop/restart, so always read the current one off the container card rather than hardcoding it.
Must serve plain HTTP — port 80 by default, or set WORKLOAD_PORT to match. Non-HTTP images (Postgres, Redis, Kafka, ...) need the TCP port below instead — the domain above will never work for them.
Publishes this container-internal port directly, bypassing the HTTP proxy entirely — connect with the service's own client (redis-cli, psql, ...), not a browser. Docker driver only. The assigned port can change after a stop/restart — always check the address shown on the card, don't hardcode it.
Some images (Postgres, MySQL, ...) refuse to boot at all without a credential passed this way. Values are sent once at launch — not visible again after, so save them if you'll need them.
Loading containers…