Cloud API (for LLMs)
Also known as: model API · inference API · API access
A cloud API lets you use a model over the internet: you send a text (and a key), the provider runs the model on its GPUs, and you get the response back — billed per token.
- Zero setup: no GPU, no download, no weights.
- Scale on demand; the provider handles uptime and capacity.
- Your text leaves your machine — the trade-off for zero ops.
Cloud vs. local in one line: the cloud gives you the biggest models with zero hardware — at the cost of per-token fees, latency, and sending your data off-site. Local inverts every one of those. Both are legit; the right choice depends on what you're optimizing.