Installation
This guide covers installing chaoscypher-core as a standalone Python library for use in your own applications.
Requirements
- Python 3.14 or later
- SQLite 3.35+ (included with Python on most platforms)
Install from PyPI
pip install chaoscypher-core
This installs the core library with all production dependencies, including:
- SQLModel and Pydantic for data modeling
- sqlite-vec for vector similarity search
- LangChain ecosystem for LLM orchestration and workflows
- LLM provider SDKs (Ollama, OpenAI, Anthropic, Gemini)
- Document processing (PDF, OCR, audio transcription)
- Content normalization (encoding fixes, deduplication)
Development Extras
For testing, linting, and type checking:
pip install chaoscypher-core[dev]
The [dev] extra adds pytest, ruff, mypy, pre-commit, pip-audit, and related tooling.
What's Included
chaoscypher-core is the framework-agnostic brain of Chaos Cypher. It provides:
| Capability | Description |
|---|---|
| Graph CRUD | Create, read, update, delete nodes, edges, and templates |
| Storage adapters | SQLite adapter with full protocol coverage |
| Entity extraction | LLM-powered extraction from documents |
| RAG pipeline | Chunking, embedding, and vector search |
| Workflow engine | LangGraph-based workflow execution |
| Graph analytics | Centrality, clustering, path analysis |
| Search | Keyword, vector, semantic, and hybrid search |
| Quality scoring | Entity and relationship quality metrics |
What Requires the Full Stack
Some features need the complete Chaos Cypher platform (Cortex API + Neuron workers + Valkey + Docker):
- Background job queue -- async processing via Valkey
- Web UI -- React interface for graph exploration
- REST API -- FastAPI endpoints with authentication
- Multi-user access -- concurrent access with proper locking
- Queue monitoring -- real-time job status dashboard
See the Getting Started guide for deploying the full platform.
Verify the Installation
import chaoscypher_core
print(chaoscypher_core.__version__)
Next Steps
- Quick Start -- build a knowledge graph in under 20 lines
- Core Concepts -- understand the architecture