Koda Docs

Commands

Command	Description
`koda pull <model>`	Download a model from HuggingFace
`koda list`	Show downloaded models
`koda list --available`	Show all pullable models
`koda run <model>`	Start an interactive chat session
`koda run <model> --system "..."`	Chat with a custom system prompt
`koda run <model> --ctx 8192`	Set the context window size
`koda run --kodafile Kodafile`	Run with a Kodafile config
`koda serve`	Start the API server on :11434
`koda serve --host 0.0.0.0 --port 8080`	Custom host and port
`koda ask <question>`	Ask a terminal question — reads your directory and shell history
`koda update`	Update koda to the latest version
`koda create`	Generate a Kodafile template
`koda version`	Show Koda version

koda pull

Download a model from HuggingFace. All models use Q4_K_M quantization.

koda pull llama3.2
koda pull mistral

koda run

Start an interactive chat session with a model.

# Basic chat
koda run llama3.2

# With a custom system prompt
koda run llama3.2 --system "You are a Python expert."

# With a larger context window
koda run llama3.2 --ctx 8192

# Using a Kodafile
koda run --kodafile Kodafile

Type /bye to exit.

koda serve

Start the API server. It implements both Ollama and OpenAI protocols.

# Default (localhost:11434)
koda serve

# Custom host and port
koda serve --host 0.0.0.0 --port 8080

koda ask

Ask a question directly from your terminal. Koda reads your current directory and shell history for context.

koda ask how do I generate a ssh key
koda ask why is port 3000 already in use
koda ask how do I undo the last git commit

After install, the ask shortcut is also available directly in your shell:

ask how do I generate a ssh key

koda list

Show downloaded models, or all available models.

koda list
koda list --available

CLI Reference