Installation

Requirements

• Python 3.10 or newer
• Network access for literature APIs (OpenAlex, Unpaywall) when not in dry-run
• Optional: Ollama or another OpenAI-compatible LLM server
• Optional: NVIDIA GPU / Apple MPS for faster PDF → markdown (Marker)

Create a virtual environment

macOS / LinuxWindows

git clone https://github.com/MineralsCloud/litcurate.git
cd litcurate
python3 -m venv .venv
source .venv/bin/activate

git clone https://github.com/MineralsCloud/litcurate.git
cd litcurate
python -m venv .venv
.venv\Scripts\activate

Install LitCurate

pip install -U pip
pip install -e ".[dev,openalex]"

Extra	Purpose
openalex	OpenAlex client (pyalex)
dev	pytest, ruff
docs	MkDocs Material (this site)

Verify:

litcurate --help
litcurate stages

Optional: local LLM (Ollama)

ollama pull qwen3:14b
ollama serve

Set OPENAI_API_KEY to any non-empty value in .env (for example ollama). Configure provider / base_url / model in your YAML — see LLM providers.

Check the dry-run

litcurate run configs/config.yaml

Expected

The pipeline completes using sample payloads when dry_run: true in the config. No API keys required for this smoke test.

To run the real pipeline, open configs/config.yaml and change:

dry_run: false

Then run the same command:

litcurate run configs/config.yaml

With dry_run: false, LitCurate performs live searches, network downloads, and configured LLM calls. Set the required environment variables before running.

Next: First steps.