Ship once and Meridian runs your code in every region you choose, close to the people using it.

Capacity grows with each spike in traffic and falls back to zero when it passes. You pay for the seconds you run.

Pick CPUs or GPUs per function and have them in seconds, with no quotas to file and nothing to reserve.

Every machine is probed without pause. Faulty nodes are drained and replaced before a request can reach them.