Self-Hosted RAG Chatbot for Your Docs.
RAGussy is a self-hosted Retrieval-Augmented Generation chatbot optimized for Markdown documentation and forum discussions. Upload your docs, configure your vector store, and get an AI-powered Q&A system with source citations. In relaunch terms: this is one of the clearest GitHub-first utility projects for turning a pile of docs into something you can actually interrogate.
Browser-based configuration panel with a setup wizard for API keys, model selection, and vector store settings. No YAML editing required.
Drag-and-drop uploads, zip file support, and bulk ingestion. Handles Markdown files and threaded forum discussions with full context preservation.
Uses Qdrant for vector storage with collection management through the UI. Supports text-embedding-3-small/large and BAAI/BGE-m3 embedding models.
AI-powered Q&A interface that answers questions about your docs and cites the specific source documents used to generate each response.
Optional Discord bot with slash commands. Deploy RAGussy as a support bot in your Discord server — members can ask questions and get sourced answers.
Works with OpenAI, OpenRouter, and any OpenAI-compatible API endpoint. Use whatever LLM provider fits your needs and budget.
RAGussy ships with Docker Compose for one-command deployment, plus automated install scripts for Linux, macOS, and Windows. The setup wizard walks you through API key configuration and vector store initialization on first run, which makes it a strong fit for teams that want self-hosted retrieval without building the whole stack from scratch.
git clone https://github.com/mojomast/ragussy.git cd ragussy docker compose up -d