This platform enables the creation and management of RAG-native knowledge graphs, targeting developers and researchers who need to build dynamic, graph-enabled AI workflows. It offers modular graph construction, flexible data ingestion, and rule-based entity resolution, facilitating the integration of structured and unstructured data for scalable AI applications.

How It Works

The studio leverages a NoSQL database backend (initially MongoDB) for flexible and scalable storage of complex relationships. It supports RAG-native graph construction, allowing for the creation of knowledge graphs that can be queried and utilized by AI models. The architecture is API-first, with a Python SDK for programmatic interaction, enabling seamless integration into existing AI pipelines.

Quick Start & Requirements

• Installation: pip install . (editable install: pip install -e .[dev,docs])
• Prerequisites: Python 3.10+, OpenAI API key, MongoDB account (M10+ recommended).
• Setup: Clone repo, configure .env with API keys and MongoDB credentials, run python admin.py setup-collections and python admin.py create-user from src/whyhow_api/cli/.
• API Launch: uvicorn src.whyhow_api.main:app
• SDK Usage: pip install whyhow, then instantiate WhyHow(api_key='...', base_url="http://localhost:8000").
• Docker: Build with docker build --platform=linux/amd64 -t kg_engine:v1 . and run with docker run -it --rm -p $OUTSIDE_PORT:8000 kg_engine:v1.
• Docs: Swagger UI available at http://localhost:8000/docs.

Highlighted Details

• Supports RAG-native knowledge graphs.
• Features rule-based entity resolution.
• API-first design with a Python SDK.
• Modular graph construction and flexible data ingestion.

Maintenance & Community

The project is from whyhow-ai. Further community or roadmap details are not explicitly provided in the README.

Licensing & Compatibility

The README does not specify a license. Compatibility for commercial use or closed-source linking is not detailed.

Limitations & Caveats

The project is described as being built on top of NoSQL, with aims for database agnosticism, but currently requires MongoDB for setup. Performance recommendations suggest M10+ MongoDB clusters, indicating potential resource requirements for optimal operation.