Local AI · Your choice

Your models. Your device.

Nimbus8 makes open models feel at home: find one that fits, download it once, and use it privately across the app. Change models whenever your work changes.

Browse & discover

Explore compatible models from Hugging Face without leaving Nimbus8. Search by name, compare download sizes, and see which choices are a comfortable fit for your device before you install.

Install & manage

Downloads can resume after an interruption. Once a model is ready, it appears anywhere in Nimbus8 that can use it. Switching takes a tap, and removing a model gives the storage back immediately.

Nimbus8 remembers where each model came from and verifies the downloaded file, so the model you return to is the model you installed.

Gated models

Some model creators ask you to accept a license before downloading. After you do that on Hugging Face, you can connect your account in Nimbus8. Your access credential stays protected on your device and is used only when Hugging Face needs to authorize a download.

Bring your own

Already have a compatible model file? Import it through Files or AirDrop. Nimbus8 checks it, measures how it runs on your device, and makes it available alongside your other local models. No Hugging Face account is required.

And models from Nimbus Labs

Nimbus-2B, Nimbus-4B, and Nimbus-9B v2.1 are public now: a family of compact coding models designed for consumer hardware, with full-precision weights, GGUF downloads, model cards, manifests, and checksums.

Meet Nimbus Models and see the verified results →