MESkit public website

Getting Started

Set up MESkit locally in under 10 minutes. You need Node.js, a Supabase project, and one LLM API key.

Last updated: March 7, 2026

Summary

What you need and what you get.

Summary

MESkit is a self-hosted MES that runs on your machine. Clone the repo, create a free Supabase project for the database, add an LLM API key for the AI agents, and start the dev server. After setup, you can build a shop floor and talk to the Operator Assistant immediately.

Prerequisites

What to have ready before you start.

Node.js 18+ — check with node --version
npm — included with Node.js
A Supabase account — free at supabase.com
An LLM API key — one of:
- Google Gemini (default, free tier available)
- OpenAI
- Anthropic

Step 1 — Clone and install

Get the code on your machine.

git clone https://github.com/meskit-cloud/meskit.git
cd meskit
npm install

Step 2 — Set up Supabase

Create the database and auth backend.

Go to supabase.com/dashboard and create a new project.
Once the project is ready, go to Settings → API and copy:
- Project URL → NEXT_PUBLIC_SUPABASE_URL
- anon/public key → NEXT_PUBLIC_SUPABASE_ANON_KEY
Open the SQL Editor in Supabase and run the contents of supabase/migrations/001_isa95_schema.sql. This creates all 15 ISA-95 tables with RLS policies, enums, indexes, and Realtime publications.

Step 3 — Configure environment

Connect MESkit to your Supabase project and LLM provider.

cp .env.local.example .env.local

Open .env.local and fill in:

# Supabase
NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-key

# LLM Provider: "google" (default) | "openai" | "anthropic"
LLM_PROVIDER=google

# Google Gemini (default)
GEMINI_API_KEY=your-gemini-api-key

# Or OpenAI (set LLM_PROVIDER=openai)
# OPENAI_API_KEY=your-openai-api-key

# Or Anthropic (set LLM_PROVIDER=anthropic)
# ANTHROPIC_API_KEY=your-anthropic-api-key

Step 4 — Start the dev server

Run MESkit locally.

npm run dev

Open localhost:3000, sign up for an account, and you are in.

Step 5 — Talk to the Operator Assistant

Your first interaction.

The chat panel is on the right side of the screen. The Operator Assistant is active by default. Try these commands to build your first shop floor:

"Create a line called Assembly"
"Add 3 workstations to Assembly"
"List my lines"
"What workstations does Assembly have?"

Every command calls the same tool layer as the UI buttons. The agent validates input, executes against Supabase, and streams the result back.

Key facts

Quick reference.

Key facts

MESkit runs locally with Node.js, Supabase, and one LLM API key.
The agent runtime supports Google Gemini, OpenAI, and Anthropic — you choose.
All 15 ISA-95 tables are created by a single migration file.
The Operator Assistant is available immediately after setup — talk to it in the chat panel.
MESkit is MIT licensed — no vendor lock-in, no usage fees from MESkit itself.

What is available today

Current milestone status.

MESkit is at M2 — Build Mode + Operator Assistant. Here is what works right now:

Build Mode — Create and manage lines, workstations, and machines via the Operator Assistant
Chat panel — Streaming responses from the agent runtime
Auth — Signup, login, logout with protected routes
ISA-95 schema — All 15 tables active with row-level security

Configure Mode (M3), Run Mode (M4), Monitor Mode (M5), and MQTT integration (M6) are planned. See the roadmap for details.

Common questions

Answers to what comes up first.

Mini FAQ

Do I need a paid Supabase plan?

No. The free tier includes Postgres, Auth, and Realtime — everything MESkit needs. You only pay for your LLM API usage.

Which LLM provider should I use?

Any of the three works. Google Gemini is the default because it offers a generous free tier. Set LLM_PROVIDER in your .env.local to switch.

Can I deploy MESkit to production?

MESkit is in active development (M2 in progress). You can deploy the frontend to Vercel and use Supabase Cloud for the backend, but expect breaking changes between milestones.