Skip to main content

​
What is Self-Hosting?

Kortix is fully open-source and can be self-hosted on your own infrastructure. This gives you complete control over your data, customization options, and deployment environment.

​
Platform Architecture

Kortix consists of four main components that work together:

​
Backend API

Python/FastAPI service that powers the agent platform with:
  • REST endpoints for agent management
  • Thread and conversation management
  • Agent orchestration and workflow execution
  • LLM integration via LiteLLM (Anthropic, OpenAI, etc.)
  • Extensible tool system
  • Real-time streaming responses

​
Frontend Dashboard

Next.js/React application providing:
  • Chat interfaces for agent interaction
  • Agent configuration dashboards
  • Workflow builders
  • Monitoring and analytics tools
  • Deployment controls

​
Agent Runtime

Isolated execution environments for each agent featuring:
  • Docker-based sandboxing
  • Browser automation capabilities
  • Code interpreter
  • File system access
  • Tool integration
  • Security isolation

​
Database & Storage

Supabase-powered data layer handling:
  • User authentication and management
  • Agent configurations
  • Conversation history
  • File storage
  • Workflow state
  • Real-time subscriptions

​
Deployment Options

Kortix supports multiple deployment methods:

Docker Compose

Quickest way to get started. All services run in containers with automatic orchestration.

Manual Setup

Run services natively for development or customization. Gives you full control over each component.

AWS ECS/EKS

Production-grade deployment with auto-scaling, monitoring, and high availability.

Lightsail

Simple cloud deployment for small to medium workloads.

​
Prerequisites

Before self-hosting Kortix, you’ll need:

​
Required Services

  • Supabase Account - For authentication and database (cloud or self-hosted)
  • Daytona API Key - For agent execution environments
  • Composio API Key - For tool integrations
  • LLM Provider - At least one of: Anthropic, OpenAI, AWS Bedrock, or OpenRouter

​
System Requirements

For Docker Setup:
  • Git
  • Docker (with Docker Compose)
  • 4GB+ RAM
  • 10GB+ disk space
For Manual Setup:
  • Git
  • Node.js 18+ and pnpm
  • Python 3.11+ and uv
  • Docker (for Redis)
  • 8GB+ RAM
  • 20GB+ disk space
For Production Deployment:
  • AWS account (for ECS/EKS)
  • Domain name (optional, for custom URLs)
  • SSL certificate (for HTTPS)

​
Features

Self-hosted Kortix gives you access to all platform capabilities:
  • AI Agent Creation - Build custom agents with specialized tools and workflows
  • Browser Automation - Navigate websites, extract data, automate web workflows
  • File Management - Create, edit, and organize documents programmatically
  • Code Execution - Run code safely in sandboxed environments
  • API Integrations - Connect with external services via Composio
  • Multi-Model Support - Use different LLM providers for different tasks
  • Real-time Streaming - Get live responses from your agents
  • Conversation History - Full chat history and thread management

​
Quick Start Path

1

Clone Repository

git clone https://github.com/kortix-ai/suna.git
cd suna
2

Run Setup Wizard

python setup.py
The wizard guides you through all configuration steps.
3

Start Services

python start.py
Access Kortix at http://localhost:3000
The setup wizard saves your progress, so you can resume if interrupted. All sensitive data is stored securely in environment files.

​
What’s Next?

Setup Guide

Detailed setup instructions for Docker and manual deployment

Configuration

Learn about all configuration options and environment variables

Deployment

Deploy to production environments with AWS

Troubleshooting

Common issues and solutions

​
Community Support

Need help with self-hosting?
  • Discord - Join our community Discord for real-time support
  • GitHub Issues - Report bugs and request features on GitHub
  • Twitter - Follow @kortix for updates

Build docs developers (and LLMs) love

Get started for freeTalk to us