Video tutorial · The Metaverse Guy
Design OS: The Free, Open-Source AI Tool That Replaces Figma
I show a 100% free way to create high-quality designs using Design OS, even without cloud code, thanks to unslop.design.
This video was made in partnership with VYBES.dev. It carries YouTube's paid-promotion label.
- Published
- 5 January 2026
- Length
- 14 min
- Views
- 12,564 views
In this sponsored video, I demonstrate Design OS, a free and open-source AI design system that simplifies UI/UX work. For those who don’t want to use cloud code, I introduce unslop.design – a hosted alternative that makes design generation accessible without deep technical skills. I walk through the full process, from project definition to exporting a design system and building a landing page in Next.js, all while reducing AI slop.
Why Design OS Matters and Who It’s For
Design OS is an open-source AI design tool that turns prompts into a structured design process. The official GitHub repo assumes you’re comfortable with cloud code, but I know that not everyone wants to work that way. That’s why I built unslop.design, a fully hosted version that wraps the same logic in an interactive form – no terminal, no cloud code needed.
Whether you’re a non-technical founder, a designer who hates Figma, or a developer fed up with ‘vibe-coded’ mess, this approach gives you a real design pipeline. And because the whole thing is open source and can run on free AI models via OpenRouter, it costs nothing to get started.
Getting Started with unslop.design
Head over to unslop.design and sign in. The platform is completely free to use – you just need to bring your own API key. I use OpenRouter because it gives access to both free and paid models. Once you paste your key, the tool connects to the AI and is ready to go.
There’s no setup, no CLI commands, and no cloud-code subscription. Just log in, add your key, and you’re inside the project dashboard.
Defining Your Project Without the CLI
Instead of typing commands, you fill in a few fields. Write a project name and a short problem statement – what the app does and who it’s for. This replaces the usual command-line flow, so anyone can describe their idea.
Hit generate and the AI creates a product overview based on that statement. It breaks down the core functionality, target users, and high-level goals. From there, I can refine it or move straight to the roadmap.
AI-Generated Design System: Colours, Typography, and Layout
The real power of Design OS is how it sequences the design work. After the product overview and roadmap, it generates a data model, then a complete design system. This includes a colour palette and typography scale, plus an app shell layout and a list of sections or modules the app will need.
You can step through each generation one by one, review the output, and tweak the prompts. Once you’re happy, the tool packages everything into exportable design system files.
Exporting the Design and Creating a Next.js Project
The export gives you a folder with a design system JSON and token files. The structure is clean and ready to drop into a front-end project. I take those files into a fresh Next.js app – no messy copying, just a straightforward import.
With the design system in place, the app already knows about colours, spacing, and typography. That means any AI coding tool you use now has a source of truth for the visual language.
Prompting the AI to Build a Landing Page
In my Next.js project, I give a simple prompt: ‘Build a landing page using the imported design system.’ Because the design tokens are already defined, the AI doesn’t invent random styles. It sticks to the palette, uses the correct typography, and follows the layout sections from the export.
The result is a polished, on-brand landing page that would otherwise take hours in Figma and front-end code. And the whole process – from problem statement to live page – takes minutes, not days.
Why Design Systems Kill AI Slop
Without a design system, AI coding tools tend to slap together whatever Tailwind classes feel right – hence ‘slop’. Design OS forces the AI to work within predefined variables, so every generated component looks and feels consistent.
I tested this with both free and paid OpenRouter models. The free models already produce great results; paid models add polish but aren’t necessary to see the benefit. The key is having a system in place before you write a single line of code.
Tools used in this video
- Design OS
- unslop.design
- OpenRouter
- Next.js
Questions
- What is Design OS?
- Design OS is an open-source AI design process that helps you generate product overviews, roadmaps, data models, and a full design system. It works best with cloud code, but you can also use the hosted version at unslop.design if you prefer a visual, no-code interface.
- Do I need to know cloud code to use this?
- No. The video shows unslop.design, which replaces cloud code with an interactive form. You just describe your project and the AI generates the design assets step by step.
- Is unslop.design completely free?
- Yes, the platform itself is free. You provide your own OpenRouter API key and can choose free AI models, so the entire workflow can cost nothing.
- What kind of output can I expect?
- You get a product overview, roadmap, data model, a design system with colours and typography, an app shell layout, and a list of app sections. These can be exported as files and used directly in a Next.js project.
Chapters
- 00:00 100% Free & Open-Source AI Design OS
- 00:23 Turning Design OS into a Web App (No Cloud Code)
- 01:10 Sponsor: VYBES.dev
- 02:46 Introducing unslop.design (Hosted Design OS)
- 03:14 Signing In & Why It’s Free
- 03:37 Adding an OpenRouter API Key (Free Models)
- 04:28 Getting Started with a Project Folder
- 05:19 Interactive Form vs CLI (Non-Technical Flow)
- 05:27 Defining the Project & Problem Statement
- 06:12 Generating the Product Overview
- 06:26 Creating the Product Roadmap
- 07:36 Generating the Data Model with AI
- 07:56 Creating the Design System (Colors & Typography)
- 08:30 Generating the App Shell Layout
- 08:44 Defining Sections & App Modules
- 09:24 Exporting the Design System Files
- 09:48 Understanding the Generated Folder Structure
- 10:35 Creating a Next.js Project
- 11:26 Importing the Design System into the Project
- 12:21 Prompting the AI to Build a Landing Page
- 13:26 Final Landing Page Result
- 13:49 Why Design Systems Reduce AI Slop
- 13:56 Free vs Paid AI Models (Final Advice)
Video description and links
This article was written from the video's description with the help of AI. Watch the video for the full walkthrough.



