Local & on-device AI

Runs models on phones and small devices, handing off to the cloud only when stuck.

The bottom of the inference market — a 14MB model and a dependency-free C++ runtime aimed at the four in five edge devices that cost under $200, where the unit economics are flash reads and milliwatt-hours rather than tokens.

Links the computers you own into one machine big enough for a large model.

The local-inference edge — the software layer that turns hardware a team already owns into the cluster, where the buy-versus-rent decision gets settled in Thunderbolt cables rather than API bills.

A desktop app that downloads and runs open models on your machine.

The default way a normal person runs a model on their own hardware — and, as of this year, a local-first company that has started selling hosted inference, which is the direction's central trade-off happening inside one product.

The one-command way to run open models on your own computer.

The clearest live test of whether local inference can hold its ground — a 9-million-developer local runtime that has started selling cloud tokens, publishing the per-token price of both sides of that choice.