Installation¶

There are two ways to install LLM Extractinator. Docker is the recommended route — it's the least fiddly and the closest thing to "it just works." A local install is the alternative if you want to develop against the package or can't use containers.

Recommended: Docker

The Docker image bundles Python, Ollama, and the Studio into one GPU-ready container. The only thing you install is Docker itself — no Python environment to manage, no separate Ollama setup. If you just want to use the tool, start there:

mkdir -p data examples tasks output ollama_models
docker run --rm --gpus all \
  -p 127.0.0.1:8501:8501 -p 11434:11434 \
  -v $(pwd)/data:/app/data \
  -v $(pwd)/examples:/app/examples \
  -v $(pwd)/tasks:/app/tasks \
  -v $(pwd)/output:/app/output \
  -v $(pwd)/ollama_models:/root/.ollama \
  lmmasters/llm_extractinator:latest

This opens the Studio at http://127.0.0.1:8501 (drop --gpus all for CPU-only). See the Docker guide for the full walkthrough, including Windows/PowerShell.

The rest of this page covers the local install.

Local install¶

Requirements¶

Python 3.10+ (3.11 recommended)
A running Ollama instance (the local LLM backend)
(Optional) Conda or venv to keep things isolated

1. Create an environment¶

conda create -n llm_extractinator python=3.11
conda activate llm_extractinator

(Or python -m venv venv && source venv/bin/activate.)

2. Install Ollama¶

Linux:

curl -fsSL https://ollama.com/install.sh | sh

Windows / macOS: download the installer from ollama.com/download.

Ollama runs a small background service. Make sure it's running before you extract — on desktop, launching the Ollama app is enough; on a headless Linux box, ollama serve starts it.

You don't need to pull models yourself

When you run a task, LLM Extractinator asks Ollama for the model you named and pulls it automatically on first use. (The exception is when you point at an externally managed server with --ollama_host — then the model must already be present there.)

3. Install the package¶

From PyPI:

pip install llm_extractinator

Or from source (handy if you want to hack on it):

git clone https://github.com/DIAGNijmegen/llm_extractinator.git
cd llm_extractinator
pip install -e .

Installing adds three commands to your environment:

Command	What it does
`launch-extractinator`	Opens the Studio (Streamlit app)
`build-parser`	Opens just the Output Schema Builder
`extractinate`	Runs a task from the terminal

4. Verify¶

Check the CLI is available:

extractinate --help

Then head to the Quickstart for a complete first run, or launch the Studio to explore:

launch-extractinator

Running without a GPU? See CPU-only hardware.