Getting Started

A small language model framework for local inference and fine-tuning.

Best Performance

For best performance, use an Apple Silicon Mac.

Installation

Quick Install (macOS / Linux)

Use the following curl command to install the framework it will install the service, update the config and configure the TUI , this is the best way to quickly get started:

curl -fsSL https://smartloop.ai/install | bash

Homebrew (macOS / Linux)

If you are more used to homebrew , we have an option for that too, it will install the artifacts but wont start the service , follow the instruction to do so:

brew tap smartloop-ai/smartloop
brew install smartloop

Upgrade via Homebrew

brew update && brew upgrade smartloop

Uninstall

# If installed via curl
curl -fsSL https://smartloop.ai/uninstall | bash

# If installed via Homebrew
brew uninstall smartloop
brew untap smartloop-ai/smartloop

Configuration

Initialize a new project with your developer token:

slp init -t <developer_token>

You can obtain your access token from https://app.smartloop.ai/developer.

Usage

# View available commands
slp --help

Project Management

slp projects create <name>
slp projects list
slp projects switch <name>
slp status

Add a Document

Upload a document for training:

slp add document.pdf

Prompting

Start the interactive chat:

slp run

For a non-TUI experience:

slp run --no-tui

Server Management

SLP includes a background API server compatible with OpenAI's chat completion format.

slp server start
slp server stop
slp server status

On macOS (Homebrew):

brew services start smartloop
brew services stop smartloop

On Linux/WSL (systemd user service):

systemctl --user start smartloop
systemctl --user stop smartloop
systemctl --user status smartloop

Supported Models

Model	Source	Size
gemma3-1b	google/gemma-3-1b-it	1B
gemma3-4b	google/gemma-3-4b-it	4B
llama3-1b	meta-llama/Llama-3.2-1B-Instruct	1B
llama3-3b	meta-llama/Llama-3.2-3B-Instruct	3B
phi4-mini	microsoft/phi-4-mini	4B

Requirements

• macOS (Apple Silicon) or Linux (x86_64)
• Python 3.11 or later (installed automatically)
• CMake (installed automatically via Homebrew)