Serving
Always warm
What is always warm?
An always-warm endpoint keeps at least one replica resident at all times, so no request ever pays a cold start. It bills continuously whether or not requests arrive, which is the honest description of reserved capacity.
It is the correct choice whenever a human is waiting on the response. Predictable latency is worth more than the saving from releasing capacity between requests.
The cost is straightforward to reason about: hourly rate multiplied by hours resident, independent of traffic. That predictability is itself a feature — the bill does not surprise you when usage spikes.
Related terms
Start with the free tier
A magic link creates your account, your tenant and your first API key. No card until you ask for compute.