Move 200GB between clouds with two curls

You have a 200GB Postgres backup on a server in Frankfurt. You need it on a box in Singapore. The standard playbook: upload to S3, wait, download from S3, wait, pay twice for egress.

You skip the middleman. On the source: pg_dump db | gzip | curl -T - https://pipe.containers.hoody.com/api/v1/pipe/migration. On the destination: curl https://pipe.../api/v1/pipe/migration | gunzip | psql newdb. Bytes flow Frankfurt → Hoody → Singapore in one continuous stream. Nothing lands on disk along the way. The pipe holds zero bytes; it's a router, not a bucket.

A spectator on a third laptop opens ?progress and watches an SSE feed of bytes-per-second and ETA, without consuming a receiver slot. You're not paying S3 for storage you'll delete tomorrow. You're paying for the seconds the bytes were in flight, and that's it.