Quickstart

Get Akio running in under a minute.

1. Pull a model

akio pull ggml-org/Qwen3-0.6B-GGUF

2. Start a session

akio run -m Qwen3-0.6B-Q4_0.gguf

That's it. Akio will load the model and drop you into an interactive chat where the agent can autonomously use tools like shell execution, file I/O, and web search.

CLI Options

akio run -m <model.gguf> -c <context_size> --ngl <gpu_layers>

Flag	Description	Default
-m	GGUF model file name	—
-c	Context window size (tokens)	model default
--ngl	Number of layers offloaded to GPU	99

Other Commands

akio list          # List cached models
akio list --all    # List all available models
akio rm <repo>     # Remove a downloaded model