Choosing a model for private, local workloads

What running a model locally actually changes

A cloud provider elsewhere in ClawAI’s catalog runs a model on its own infrastructure and charges per request; Ollama and llama.cpp instead load an open-weight model onto hardware you control, so the request never leaves it. That changes who can see the request, not what any given model is capable of — see local AI, on the model providers page, for the full mechanism rather than repeating it here.

ClawAI’s Local-Only and Privacy-First routing modes

Local-Only routing keeps every request on hardware you control, using Ollama or llama.cpp rather than any cloud provider. Privacy-First routing is a separate mode with its own priorities; both exist specifically because not every workload should default to Auto routing. Choosing between them, or pinning a specific local model under Manual Model mode, is a workload decision worth making deliberately rather than leaving to a general-purpose default.

Choosing which open-weight model to run

This page deliberately names no specific open-weight model, for the same reason the local AI provider page does not: the field moves faster than a static page can track, and a stale recommendation is worse than none. See what is local-first AI, linked below, for how to think about the trade-off between an open-weight model you run yourself and a cloud provider.

Questions people ask

Which open-weight model should I run for a private workload?

This page does not recommend one — see what is local-first AI, linked below, for how to think about the choice, since the right model depends on your hardware and task in a way a static page cannot track responsibly.

What is the difference between Local-Only and Privacy-First routing?

Local-Only keeps every request on hardware you control via Ollama or llama.cpp; Privacy-First is a separate routing mode with its own priorities. Both exist because not every workload should default to Auto routing.

Does running a model locally cost anything through ClawAI?

ClawAI does not charge a per-token rate for a locally run model the way it does for a cloud provider, since there is no cloud provider being billed — the cost is the hardware you already run it on. Confirm current plan behaviour on the pricing page.