Skip to main content
While our SaaS offering is the fastest way to get started, Corgtex’s On-Premise and Hybrid deployment modes allow you to run the intelligence layer entirely on your own infrastructure so your data never leaves your building. The simplest way to deploy the stack is using Docker Compose.

​
Prerequisites

Before starting, ensure your host machine has:
  • A modern Linux distribution (Ubuntu 22.04 LTS recommended)
  • Docker and Docker Compose installed
  • Node.js 22 (for optionally running local scripts)
  • At least 4GB of RAM and 2 vCPUs

​
The Deployment Guide

1

Clone the repository

Download the Corgtex source code onto your host server.
git clone https://github.com/corgtex/corgtex.git
cd corgtex
2

Configure environment variables

Copy the sample environment file to create your active configuration.
cp .env.example .env
Open .env in an editor and set your core variables. Pay special attention to the secrets.
# Required for production
SESSION_COOKIE_SECRET="generate_a_random_32_char_string"
APP_URL="https://your_domain.com"
See the Configuration Reference for a complete list of required and optional variables.
3

Start the stack

Run the self-host compose file. This file starts the PostgreSQL database, the Next.js Web server, and the background Worker process.
docker compose -f docker-compose.selfhost.yml up -d
4

Access the application

The web container exposes port 3000. If you are running locally:http://localhost:3000If you are deploying to a remote server, you should configure a reverse proxy (like Nginx, Traefik, or Caddy) to route external traffic to 127.0.0.1:3000 and handle TLS/SSL termination.

​
What’s included in the Self-Host bundle?

When you run docker-compose.selfhost.yml, you are launching three distinct services:
  1. postgres: A postgres:16-alpine container running the core database and (pgvector). The data is persisted locally in a Docker volume.
  2. web: The Next.js application that handles all API requests and user interface rendering.
  3. worker: The background Node.js daemon that executes long-running tasks asynchronously.