← Glossary
Glossary · AI and agent workflows
Local inference
Running a model on a machine you control, whether a laptop or a server you administer, so that the model's own computation happens there rather than on a provider's servers.
Local inference settles where the model runs. It does not by itself settle where data goes: the harness around the model may search the web, send telemetry, back files up or call a cloud model for some requests, and each of those is a separate route to check. One product's documentation distinguishes its local models from its cloud models and documents how to switch cloud features off; that is the shape of the check. Local running also has costs a zero provider charge hides: the machine, its power and its upkeep.