graphrag

Translate natural language GraphRAG descriptions into PostgreSQL schemas, indexes, and queries.

Updated Dec 25, 2025

Search Tags:#embeddings #pgvector #graph #knowledge-graph #postgres #graphrag #pgroonga

Ready to run

npx skills add https://github.com/titabash/claude-plugins --skill graphrag-titabash

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: graphrag
Source: https://github.com/titabash/claude-plugins/tree/main/graphrag-postgresql/skills/graphrag
Command: npx skills add https://github.com/titabash/claude-plugins --skill graphrag-titabash

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

GraphRAG design is complex: it translates natural language project descriptions into a GraphRAG-ready PostgreSQL design, including Entity/Edge types, schemas, and queries.

Core Features & Use Cases

Generate Entity/Edge type schemas from requirements and map them to a relational schema, vector embeddings, and search indexes.
Produce end-to-end artifacts: sql/schema.sql, sql/indexes.sql, sql/queries, and docs/prompts for extraction and summarization.
Use cases include planning a knowledge graph, building time-aware relationships, and enabling Local/Global/Hybrid search patterns.

Quick Start

Describe your GraphRAG project in natural language, and the system will produce the schema, indexes, queries, and documentation.

Frequently Asked Questions about graphrag

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I design a GraphRAG schema from a natural language project description?▼

To design a GraphRAG schema from natural language, you translate project descriptions into concrete PostgreSQL artifacts, including Entity/Edge schemas, vector embeddings via pgvector, and full-text search indexes using PGroonga.

What is the best way to build a knowledge graph schema using PostgreSQL for GraphRAG?▼

Building a knowledge graph schema in PostgreSQL involves mapping Entity and Edge types to a relational structure, applying pgvector for embeddings, and generating ready-to-use SQL templates for Local, Global, and Hybrid search workflows.

Does GraphRAG with pgvector and PGroonga support time-series relationship queries?▼

Yes, GraphRAG with pgvector and PGroonga supports optional time-series relationships, enabling time-aware queries alongside vector embeddings and full-text search within your PostgreSQL knowledge graph implementation.

How do I generate SQL schema and query templates for a GraphRAG implementation?▼

You generate SQL schema and query templates by translating natural language requirements into end-to-end PostgreSQL artifacts, outputting ready-to-use files like sql/schema.sql, sql/indexes.sql, and sql/queries with validation against initial requirements.

Can I use pgvector and PGroonga together for hybrid search in a PostgreSQL knowledge graph?▼

Yes, you can use pgvector and PGroonga together to enable Hybrid search in a PostgreSQL knowledge graph, combining vector similarity search with full-text search capabilities within your generated GraphRAG schema.

What artifacts are needed to implement Local, Global, and Hybrid search workflows in GraphRAG?▼

Implementing Local, Global, and Hybrid search workflows requires artifacts like sql/schema.sql, sql/indexes.sql, sql/queries, and documentation for extraction and summarization, all generated from your natural language project description.