Local alternatives to popular AI products

Tools that run models on hardware you control instead of requiring a cloud chat provider.

Loading catalogue…

About this collection

Why this collection exists

Popular AI products often mean sending prompts, documents, or media to a remote model provider. This collection groups tools that can run inference or AI-assisted workflows on hardware you control, such as local runtimes, private model hosts, or first-party tools that keep optional AI processing under a clear on-device or user-controlled boundary.

Privacy context

Local AI reduces prompt and document exposure to a vendor model API, but it is not automatic privacy. Local tools still involve model provenance, telemetry, account options, and what leaves the machine during updates. Compare architectures and evidence rather than treating “local” as a pass/fail badge.

Inclusion criteria

Membership is curated against the rules below. It is not an automatic filter and does not grant review or verification status.

Primary AI or model workflow can run locally, on self-hosted infrastructure, or on-device without a mandatory third-party cloud model for core use.
Listings include local model runners, private GPT-style apps, and Thryve tools that expose local or clearly scoped AI paths.
Hybrid tools may appear when a documented local path exists; pure cloud-only AI assistants are excluded.
Model downloads, optional cloud models, and update checks may still use the network; those details stay on each privacy profile.

Related collections

Explore other curated privacy shortlists and inclusion criteria.