LLiMa CLI

Use the llima CLI on Modalix to manage precompiled models and do simple runtime testing. It is useful for checking that a model loads, accepts prompts, and produces output before you integrate it with Neat Framework direct APIs or the Neat GenAI server endpoints.

Model Manager

LLiMa includes a model manager through the llima CLI. It lets you search, download, list, remove, and run precompiled models directly from the command line. Models are stored under /media/nvme/llima/models by default. Set LLIMA_MODELS_PATH to use a different models directory.

Browse available models:

modalix:~$ llima search
modalix:~$ llima search qwen

Download a model by name, without the simaai/ organization prefix:

modalix:~$ llima pull Qwen3-VL-4B-Instruct-GPTQ-a16w4

List and remove locally installed models:

modalix:~$ llima list
modalix:~$ llima rm Qwen3-VL-4B-Instruct-GPTQ-a16w4

Running LLiMa

Use llima run as a simple runtime for initial model validation on Modalix.

modalix:~$ llima run <model> [options]

Argument	Description
model	Model ID or path (e.g., Qwen3-VL-8B-Instruct-a16w4).
--stt_model_path	Path to the elf files for a Speech-to-Text model (optional).

For all available options, run llima run -h.

Examples

modalix:~$ llima run Qwen3-VL-4B-Instruct-a16w4

Interactive Commands

Once llima run starts in CLI mode, use these commands at the prompt:

Command	Description
add image <file>	Add an image to the current prompt context.
clear image	Clear all images.
set system <prompt>	Set the system prompt.
clear system	Clear the system prompt, chat history, and images.
clear history	Clear chat history and images.
print history	Print chat history.
set audio <file>	Set the audio file to transcribe as the query.
set language <lang>	Set the language string used for transcription.
set lora <name>	Use LoRA weights from a npy_files folder.
unset lora	Revert the LoRA model to the baseline model.
quit	Quit.
list command	Print available commands.
help	Print available commands.