Installation

Requirements

• Python 3.11 (recommended)
• A running Ollama instance
• (Optional) Conda or venv for isolation

Instead of installing locally, you can also run the entire application in a Docker container (see the Docker guide).

1. Create environment

conda create -n llm_extractinator python=3.11
conda activate llm_extractinator

(You can use python -m venv venv instead.)

2. Install Ollama

Linux:

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

Windows / macOS: download from ollama.com

Make sure the Ollama service is running before you try to extract.

3. Install the package

pip install llm_extractinator

Or from source:

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

Next

• go to Preparing Data to see how your CSV/JSON should look
• or run launch-extractinator to explore the Studio