Installation

From PyPI

pip install curate-ipsum

Or with uv:

uv pip install curate-ipsum

Optional extras

# Z3 SMT solver for formal verification (recommended)
pip install "curate-ipsum[verify]"

# RAG pipeline — ChromaDB + sentence-transformers
pip install "curate-ipsum[rag]"

# Graph-spectral analysis (scipy + networkx)
pip install "curate-ipsum[graph]"

# Cloud LLM synthesis (httpx)
pip install "curate-ipsum[synthesis]"

# Kuzu graph database backend
pip install "curate-ipsum[graphdb]"

# GPU-accelerated embeddings
pip install "curate-ipsum[embeddings-gpu]"

# Kitchen sink for development
pip install "curate-ipsum[dev,verify,rag,graph,synthesis,graphdb]"

Docker

The Docker image includes the full server with the all-MiniLM-L6-v2 embedding model baked in — no Python installation required.

docker pull ghcr.io/egoughnour/curate-ipsum:latest

From source (development)

git clone https://github.com/egoughnour/curate-ipsum.git
cd curate-ipsum
uv sync --extra dev --extra verify --extra rag --extra graph --extra synthesis --extra graphdb
uv run pre-commit install  # set up pre-commit hooks

MCP client configuration

Claude Desktop

Add to your claude_desktop_config.json:

Other MCP clients

Curate-Ipsum speaks the standard MCP stdio transport. Any client that supports command + args configuration can launch it with:

command: curate-ipsum

Or equivalently: uv run curate-ipsum, python -m server, or docker run -i --rm ghcr.io/egoughnour/curate-ipsum.