Summary

This project scaffolds AI agents with graph-based reasoning memory, enabling developers to rapidly generate fully-functional, domain-specific context graph applications. It targets users needing sophisticated AI with traceable, persistent knowledge, offering a complete full-stack solution in minutes, complete with Neo4j integration, custom domain support, and SaaS data import capabilities.

How It Works

The core is an interactive CLI tool that generates a FastAPI backend, a Next.js frontend with interactive graph visualization and streaming chat, and a Neo4j database schema. It leverages neo4j-agent-memory to implement a three-tiered memory architecture: short-term (conversation history), long-term (knowledge graph via POLE+O model), and reasoning memory (decision traces with provenance). This approach allows agents to reason over structured knowledge graphs with full traceability, differentiating it from simple RAG systems.

Quick Start & Requirements

• Primary Install/Run: uvx create-context-graph or npx create-context-graph.
• Prerequisites: Python 3.11+, Node.js 18+, Neo4j 5+.
• Setup: Application scaffolding in under 5 minutes.
• Links: Project repository (implied by name), Neo4j Community Forum (https://neo4j.com/community/).

Highlighted Details

• Custom Domains: Generates complete ontologies, agent tools, and schemas from plain English descriptions using an LLM.
• SaaS Data Connectors: Integrates data from services like GitHub, Slack, Jira, Notion, Gmail, Google Calendar, and Salesforce.
• Multi-Framework Support: Compatible with various agent frameworks including PydanticAI, LangGraph, CrewAI, and OpenAI Agents SDK.
• Traceable Reasoning: Implements a three-memory architecture (short-term, long-term knowledge graph, reasoning memory) for full decision provenance.
• Interactive Visualization: Features streaming chat, real-time tool call visualization, and an interactive Neo4j graph explorer.

Maintenance & Community

This project is part of Neo4j Labs, maintained by Neo4j staff and the community. It is explicitly "not officially supported." Assistance is available via GitHub Issues or the Neo4j Community Forum.

Licensing & Compatibility

• License: Apache-2.0.
• Compatibility: The Apache-2.0 license is permissive, allowing for commercial use and integration into closed-source projects.

Limitations & Caveats

The project is designated as "not officially supported," implying community-driven maintenance and potential for less formal support structures. Setup requires specific versions of Python, Node.js, and Neo4j, and LLM API keys are necessary for certain advanced features like custom domain generation.