local ai

Models can make mistakes. Everything runs locally after download.

Add a GGUF model

Use one GGUF file or all parts of a split GGUF model.

Copy GGUF files onto this device

Models above 2 GB must be split first with llama-gguf-split --split-max-size 512M. Select every numbered part together. Files are stored privately by this browser.

Display name

Context Native limit KV cache

Choose GGUF file(s)

or use a direct GGUF URL