Self-Host Open Generative AI with Docker or Node.js

Open Generative AI is designed to run on any server or workstation you control. There is no licence server, no telemetry, and no server-side credential to configure — your Muapi API key lives only in the browser of whoever uses the app. You can be up and running in minutes with either Docker Compose or a plain Node.js production build.

Prerequisites

Docker
Node.js

Docker Engine 24+ and Docker Compose v2+
A Muapi.ai API key (entered by each user in the app UI — not required at deployment time)

Deploy

Docker
Node.js

The repository ships a production-ready docker-compose.yml and a multi-stage Dockerfile. The Dockerfile uses Node 20 Alpine, builds all workspace packages, compiles the Next.js app, and produces a minimal runner image.docker-compose.yml

services:
  open-generative-ai:
    build: .
    container_name: open-generative-ai
    ports:
      - "3001:3000"
    environment:
      - NODE_ENV=production
    restart: unless-stopped

The app listens on port 3000 inside the container and is mapped to port 3001 on the host. To use a different host port, change the left side of the port mapping (e.g. "8080:3000").Build and start

git clone --recurse-submodules https://github.com/Anil-matcha/Open-Generative-AI.git
cd Open-Generative-AI
docker compose up --build -d

The first build downloads the Node base image, installs dependencies, and compiles all packages — this takes a few minutes. Subsequent builds reuse Docker layer cache and are much faster.Once running, open http://localhost:3001 in your browser.Stop the container

docker compose down

View logs

docker compose logs -f open-generative-ai

Clone the repository with submodulesThe workflow builder, AI agents, and design agent packages are git submodules. You must clone with --recurse-submodules or initialise the submodules separately:

git clone --recurse-submodules https://github.com/Anil-matcha/Open-Generative-AI.git
cd Open-Generative-AI

# If you already cloned without --recurse-submodules:
git submodule update --init --recursive

Install and build workspace packagesThe setup script initialises submodules, installs all npm dependencies, and builds the four workspace packages in the correct order (workflow-builder → ai-agent → design-agent → studio):

npm run setup

Running npm install alone is not sufficient. The workspace packages must be built before the Next.js server can start. Always use npm run setup on a fresh clone.

Build and start the production server

npm run build
npm run start

The server starts on port 3000 by default. Open http://localhost:3000 in your browser.Use a custom port

PORT=8080 npm run start

Upgrading

Docker
Node.js

Pull the latest source and rebuild the image:

git pull
docker compose up --build -d

Docker Compose will rebuild the image with the new source, stop the old container, and start a replacement. Downtime is typically under 30 seconds.

Pull the latest source, update submodules, rebuild packages, and restart:

git pull
git submodule update --recursive
npm run setup
npm run build
npm run start

Notes

No server-side API key required. Each user enters their own Muapi.ai API key directly in the app UI when they first open it. The key is saved in their browser’s localStorage and is sent in the x-api-key header on every generation request. You do not need to set any API key environment variable on the server.

Local inference is not available in self-hosted web deployments. The sd.cpp engine and Wan2GP integration are features of the Electron desktop app only. Users of your self-hosted instance will use Muapi.ai cloud APIs for all generation requests. If you need local inference, use the desktop app instead.

Get Started

Studios

Local Inference

Configuration & Architecture

Self-Host Open Generative AI with Docker or Node.js

Prerequisites

Deploy

Upgrading

Notes

Build docs developers (and LLMs) love

Get Started

Studios

Local Inference

Configuration & Architecture

Documentation Index

​Prerequisites

​Deploy

​Upgrading

​Notes

Build docs developers (and LLMs) love

Prerequisites

Deploy

Upgrading

Notes