HiGoalVita is a production-ready, full-stack AI RAG platform designed to ground LLM responses in proprietary enterprise data. It offers a modular and extensible suite for businesses needing to build conversational knowledge assistants, targeting developers and organizations seeking to leverage their internal data with AI.

How It Works

HiGoalVita employs a layered architecture, ingesting data from various sources into a unified relational-graph database using proprietary graph-network algorithms. This approach enables context-aware answers, segmentation, network analysis, and actionable workflows via multi-agents, combining RAG, intelligent agents, and graph-based indexing for a deep understanding of entities and relationships.

Quick Start & Requirements

• Backend Only: Clone repo, poetry install, then higoalcore index and higoalcore query --query "...".
• Complete Suite: Clone repo, poetry install, pull and run Redis container (docker pull redis:latest, docker run -d --name some-redis -p 6379:6379 redis:latest), start backend (uvicorn higoalengine.app.main:app), and run frontend (cd vue, npm run serve).
• Prerequisites: Python 3.10-3.12, Poetry, Docker (for full suite). Redis is required for the complete suite.
• Docs: installation_guide.md

Highlighted Details

• Production-grade backend built on FastAPI + Gunicorn, containerized with Docker/Docker-Compose, and horizontally scalable.
• Modular architecture with configurable components for databases (SQLite, MySQL, OceanBase), vector stores (LanceDB, OceanBase), caching, and LLM providers (OpenAI, DeepSeek, Qianwen).
• Features agents for NL2SQL (database querying, analytics, visualization) and domain-specific workflows (planned for v3.0).
• Includes a Web UI for interactive Q&A and document management, with persistent logging for auditability.

Maintenance & Community

The project is developed and maintained by HiGoalV Corporation. Roadmap is released incrementally, inviting community contributions. Contact zhuyang@higoall.com for commercial inquiries.

Licensing & Compatibility

Licensed under the Apache License 2.0. Incorporates unmodified code from Microsoft's graphrag (MIT License). Compatible with commercial use.

Limitations & Caveats

The project is released incrementally, with features like graph database support, role-based access control, and advanced agents marked as "coming soon" (v2.0/v3.0). Detailed documentation for all modules is still being released.