Kolosal AI Retrieval Management System

⚡ All-in-one, fully-local LLM retrieval orchestration: parse, embed, search, deploy, done.

Overview

Kolosal AI Retrieval Management System is an integrated, all-in-one solution combining a powerful LLM inference engine, embedding engine, robust document parser, high-performance vector database, built-in internet search capability, and a centralized management dashboard to seamlessly orchestrate all components. This unified platform fulfills over 90% of your LLM integration requirements while ensuring complete local deployment, privacy, and data control.

Features

• MCP (Model Context Protocol)

• RAG (Retrieval-Augmented Generation)

• LLM & Embeddings Inference

• AI Agents

• Internet Search

• Dashboard

Built On Top Of

This system integrates the following components:

• Kolosal AI Inference Engine - LLM inference server (docker branch)
• Kolosal RMS Dashboard - Web-based management interface
• Kolosal RMS MarkItDown - Document parsing service
• Qdrant Vector Database - High-performance vector database
• SearXNG Docker - Privacy-respecting search engine

Getting Started

Quick Docker Setup (Recommended)

The easiest way to get started - just 2 commands:

# 1. Clone this repository
git clone https://github.com/KolosalAI/Retrieval-Management-System.git
cd Retrieval-Management-System

# 2. Start all services
docker-compose up -d

All services will be automatically built from their respective GitHub repositories. No setup scripts or submodules needed!

Access Services:

• Dashboard: http://localhost:3000
• Kolosal Server (AI Inference): http://localhost:8084
• Qdrant Vector DB: http://localhost:6333
• SearXNG Search: http://localhost:8090
• MarkItDown API: http://localhost:8001

For detailed Docker setup instructions, see DOCKER_SETUP.md.

Configuration

Contributing

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.