Install and start

Prepare Python 3.11 or later and a running Docker installation. Start in an empty working directory.

1. Install and run

Run these commands in order:

python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade --pre 'decision-layer[stats,mcp]'
decision-layer demo

This installs the admin UI and analysis libraries, prepares sample data and starts the server. The first run may take a few minutes to download the required files.

2. Explore the UI and files

Open the admin UI to explore the prepared revenue analysis. Inspect the queried data and results in Runs.

Editable Skills and Python Methods are created in your working directory:

workspace/
├── decision-layer.yaml
└── sample_shop/methods/
    ├── metric_lookup/SKILL.md
    ├── drilldown/SKILL.md
    ├── ols/method.py
    └── matching/method.py

SKILL.md contains analysis instructions for AI; method.py contains the Python calculation. See Method projects to edit or add definitions.

3. Connect your AI client

Keep the server running and follow MCP setup to connect your AI client. Then try asking:

Find the highest-revenue department in February 2026 and explain its revenue by region.

Stop and restart

Press Ctrl+C in the terminal to stop. Code, data and execution history are kept. Run decision-layer demo from the same project directory to start again.

To analyze your own data, see Connect Cube.