LightningRAG/LightningRAG
LightningRAG is a full-stack Vue + Gin starter with a decoupled frontend and backend, plus built-in, extensible RAG (retrieval-augmented generation): knowledge bases, vector search, and integrations with many LLM and vector-store providers
What it solves
LightningRAG provides a full-stack starter for building RAG (Retrieval-Augmented Generation) applications. It simplifies the process of ingesting documents, managing knowledge bases, and deploying AI agents that can be integrated into third-party messaging platforms via webhooks.
How it works
The project uses a Vue.js frontend and a Gin (Go) backend. It implements a complete RAG pipeline: documents are parsed, chunked, and embedded into pluggable vector stores (like pgvector or Elasticsearch). Users can then create AI agents using a visual canvas flow editor to define multi-step retrieval and LLM logic. These agents can be connected to external channels like Slack, Discord, and Teams through webhook connectors.
Who it’s for
It is designed for developers, designers, and product managers who need a ready-made framework to build and deploy production-ready RAG apps with built-in user management, authority control, and agent orchestration.
Highlights
- Visual Agent Orchestration: A canvas-based flow editor for creating complex, branching AI automation.
- Pluggable Architecture: Supports various LLMs, embedding models, rerankers, and vector stores.
- Multi-Channel Integration: Built-in webhook connectors for platforms including Feishu, DingTalk, WeChat, Slack, and Telegram.
- Full-Stack Starter: Includes administrative tools like JWT authentication, Casbin-based role management, and a form builder.