Agents around the clock, no machine to keep running
Put agents on Cloud slots YATFA operates. A standard slot is free to hold, a quiet fleet costs nothing, and you pay only for measured use above a free allowance. Build a month below.
What a month costs
Measured, never reserved
Every agent is billed from its measured hourly peaks — not from the size of the slot. Memory up to 1 GB and a slice of CPU each hour are free. And if a sample goes missing, the hour can only cost less, never more.
Memory above 1 GB — hourly measured peak, per agent — is $0.014 / GB·h. CPU above the allowance — 0.05 core-hours free every agent-hour — is $0.01 / core·h.
Honest capacity, even when it’s busy
Slots are burstable, not reserved. When a host runs short, it degrades in a managed way instead of failing at random. Start the queued tickets one by one and read what happens.
Bigger slots when builds need them
Each doubling adds $5 a month, and the price per gigabyte falls as you go. Usage is measured on top, exactly as on a standard slot.
A Postgres for development
An easy start for the apps your agents build and test. Not a production database, and we don’t pretend it is.