Quick Setup
Deploy FluidRAG with Docker Compose. Four images, eleven containers, one docker-compose up -d.
Getting Started with FluidRAG
FluidRAG is a hybrid Retrieval-Augmented Generation engine that combines vector similarity search, FalkorDB knowledge graph traversal, and multi-agent workflows to deliver accurate, cited answers from your documents and data sources.
Overview
Section titled “Overview”OpenAI-Compatible API
Drop-in replacement for the OpenAI chat completions endpoint. Works with any OpenAI-compatible client or frontend.
Knowledge Graph
Built on FalkorDB for graph-based relationship discovery alongside semantic vector search and code search.
Prerequisites
Section titled “Prerequisites”Before you begin, you will need:
- Docker Engine 20.10+ and Docker Compose 2.0+
- A running FalkorDB instance (external, not bundled)
- An LLM API key (OpenAI, Anthropic, Google, or any OpenAI-compatible provider)
- At least 4 GB of available RAM
Access Points
Section titled “Access Points”Once running, FluidRAG is available at:
| Service | URL | Purpose |
|---|---|---|
| Chat API | http://localhost:8000 | OpenAI-compatible API endpoint |
| Admin UI | http://localhost/config-ui | Configure agents, secrets, prompts, search settings |
| Gateway | http://localhost | Nginx reverse proxy (routes all traffic) |
| RQ Dashboard | http://localhost:9181 | Background job monitoring |
Documentation
Section titled “Documentation”| Topic | Description |
|---|---|
| Prerequisites | System requirements, external dependencies, and LLM provider setup |
| Quick Start | Clone, configure, build images, and start FluidRAG |
| Your First Query | Send a query via the API and configure secrets through the Admin UI |