Skip to main content

Quickstart

Use this path for the first successful SLMCortex run with the fewest moving parts.

It stays close to the current public product story:

  • install the base CLI
  • check the public entry point
  • run the no-model demo
  • inspect the Composer-first commands

1. Install​

On macOS and Linux, start with Homebrew:

brew tap alvarolorentedev/SLMCortex
brew install slmcortex
slmcortex --help
slmcortex-composer --help

Or install directly from the tap in one command:

brew install alvarolorentedev/SLMCortex/slmcortex

Use a source checkout only when you specifically want the repo-local demo scripts, editable development, or deeper Factory work:

python3 -m venv .venv
. .venv/bin/activate
pip install --upgrade pip
pip install -e '.[test]'

Install a real-model backend only when you need actual training or inference:

pip install -e '.[mlx]'   # macOS Apple Silicon
pip install -e '.[gguf]'  # Linux, Windows, macOS Intel, or explicit GGUF use

backend: auto uses MLX on macOS arm64/aarch64 and GGUF everywhere else.

2. Check The Public CLI​

slmcortex --help
slmcortex doctor

doctor is the quickest way to confirm the packaged workspace contract, backend visibility, and general install health.

3. Run The No-Model Demo​

DEMO_ROOT="$(mktemp -d "${TMPDIR:-/tmp}/slmcortex-demo.XXXXXX")"
python scripts/run_slmcortex_demo.py --output-root "$DEMO_ROOT"

This source-checkout demo is the fallback validation path when you want to inspect the product layers from the repository itself.

This validates the full product path without loading a real model:

  • package two checked-in adapters
  • compose one runtime bundle
  • validate the runtime
  • run inference in --dry-run
  • run the bounded agent in --dry-run

Expected outputs under $DEMO_ROOT:

python_slm/
debugging_slm/
runtime/
agent-trace.json

4. Inspect The Composer-First Path​

The default product direction is Composer-first:

slmcortex doctor
slmcortex compose-folder --help

This is the path to use when you want to point SLMCortex at a folder, compose a runtime, and export the result without manually walking through Factory commands first.

5. Source Checkout Alternative​

If you are intentionally working from a source checkout and want to see the product layers one command at a time:

DEMO_ROOT="$(mktemp -d "${TMPDIR:-/tmp}/slmcortex-demo.XXXXXX")"

slmcortex package-slm \
  --slm-id python_slm \
  --name "Python Slm" \
  --adapter-dir artifacts/adapters/python_slm \
  --train-dataset tests/fixtures/slmcortex_demo/train.jsonl \
  --eval-dataset tests/fixtures/slmcortex_demo/eval.jsonl \
  --eval-summary tests/fixtures/slmcortex_demo/eval-summary.json \
  --output "$DEMO_ROOT/python_slm"

slmcortex package-slm \
  --slm-id debugging_slm \
  --name "Debugging Slm" \
  --adapter-dir artifacts/adapters/debugging_slm \
  --train-dataset tests/fixtures/slmcortex_demo/train.jsonl \
  --eval-dataset tests/fixtures/slmcortex_demo/eval.jsonl \
  --eval-summary tests/fixtures/slmcortex_demo/eval-summary.json \
  --output "$DEMO_ROOT/debugging_slm"

slmcortex compose-slms \
  --slms "$DEMO_ROOT/python_slm,$DEMO_ROOT/debugging_slm" \
  --output "$DEMO_ROOT/runtime"

slmcortex validate-runtime --runtime "$DEMO_ROOT/runtime"

slmcortex infer \
  --runtime "$DEMO_ROOT/runtime" \
  --prompt "Fix this Python traceback" \
  --dry-run

slmcortex agent run \
  --runtime "$DEMO_ROOT/runtime" \
  --repo /path/to/local/repo \
  --task "Fix the failing answer implementation." \
  --dry-run