UI Preset MCP Server
Enables auto-configuration and validation of React components against a glassmorphic design token system using an AST correction engine. It supports managing UI presets, enforcing design uniformity in Monaco IDE, and exporting tokens as CSS, Tailwind, or JavaScript variables.
README
UI Preset MCP Server
MCP server for auto-configuring React UI against a swappable glassmorphic preset system. Integrates with your Monaco IDE to enforce design uniformity across all business builds.
Architecture
ui-preset-mcp-server/
├── src/
│ ├── index.ts # Server entry point (stdio + HTTP transports)
│ ├── constants.ts # Shared constants and paths
│ ├── types/
│ │ └── index.ts # All TypeScript types (tokens, presets, corrections)
│ ├── schemas/
│ │ └── toolSchemas.ts # Zod validation schemas for every tool
│ ├── services/
│ │ ├── presetLoader.ts # Reads + deep-merges preset files with caching
│ │ ├── sessionState.ts # Active preset session management + overrides
│ │ ├── uiCorrector.ts # AST correction engine (10 correction passes)
│ │ ├── tokenResolver.ts # Resolves {{token:x.y.z}} placeholders + export generators
│ │ └── fileWatcher.ts # Hot-reload presets on disk changes (dev mode)
│ ├── tools/
│ │ ├── presetTools.ts # load_preset, swap_template, list_presets, diff_presets, scaffold_preset
│ │ └── correctionTools.ts # autocorrect_component, validate_ui, generate_component, generate_tokens, apply_token_overrides
│ ├── routes/
│ │ └── uiRoutes.ts # REST API + Design Studio dashboard (GET /, /api/*)
│ └── ui/
│ └── dashboardHtml.ts # Embedded glassmorphic Design Studio HTML (7 sections)
└── presets/
├── glassmorphic-base/ # Core preset (all others inherit from this)
│ ├── manifest.json
│ ├── tokens.json # Full design token system
│ ├── components/
│ │ ├── shell/Sidebar.json
│ │ ├── surfaces/GlassCard.json
│ │ ├── settings/OptionGroup.json
│ │ └── navigation/NavGroup.json
│ └── layouts/DashboardLayout.json
├── client-fintech/ # Blue accent, dense spacing
├── client-saas/ # Purple accent, wider cards
└── client-dark-minimal/ # Monochrome, reduced glass intensity
MCP Tools
Preset Management
| Tool | Description |
|---|---|
load_preset |
Activate a preset bundle by ID |
swap_template |
Hot-swap active preset without restart |
list_presets |
List all available presets |
diff_presets |
Compare token/component differences between two presets |
scaffold_preset |
Generate a new preset directory from a parent |
get_session_state |
Check active preset and runtime overrides |
Correction & Generation
| Tool | Description |
|---|---|
autocorrect_component |
Auto-fix React component against active preset |
validate_ui |
Validate component and get conformance score (0–100) |
generate_component |
Generate a component from a preset template |
generate_tokens |
Export tokens as CSS vars, JS, JSON, or Tailwind config |
apply_token_overrides |
Layer runtime token overrides on the active preset |
Installation
npm install
npm run build
Usage
stdio (for Monaco IDE integration)
node dist/index.js
HTTP server
TRANSPORT=http PORT=3001 node dist/index.js
Dev mode with hot-reload
WATCH_PRESETS=true node dist/index.js
Design Studio UI
When running in HTTP mode, a glassmorphic Design Studio mini UI is served at GET /.
UI Sections
| Section | Description |
|---|---|
| Dashboard | Active preset overview with component/layout stats and color palette preview |
| Presets | Browse all presets in cards; click Load to activate any preset instantly |
| Tokens | Visual token viewer — color swatches, typography scale, blur, spacing, animation |
| Validate | Paste React code and get a conformance score (0–100) with issue list |
| Correct | Auto-correct React code against the active preset; choose context and mode |
| Export | Generate CSS custom properties, TypeScript const, JSON, or Tailwind config |
| Scaffold | Form to create a new preset from any parent with optional accent color |
The UI communicates with the server via a REST API also available at /api/*.
REST API (HTTP mode)
| Endpoint | Method | Description |
|---|---|---|
/api/presets |
GET | List all presets with metadata |
/api/presets/load |
POST | Load and activate a preset { preset_id } |
/api/session |
GET | Get active preset and override state |
/api/tokens |
GET | Get effective tokens for the active preset |
/api/tokens/export |
POST | Export tokens `{ format: 'css' |
/api/tokens/overrides |
POST | Apply runtime token overrides { overrides } |
/api/validate |
POST | Validate React code { code, include_suggestions } |
/api/correct |
POST | Autocorrect React code { code, context, dry_run } |
/api/scaffold |
POST | Create new preset { preset_id, name, description, extends, accent_color } |
Typical Workflow
1. load_preset("glassmorphic-base") # Activate base preset
2. autocorrect_component(code) # Fix component on save
3. validate_ui(code) # Get conformance score
4. generate_tokens({ format: "css" }) # Export CSS variables
5. swap_template("client-fintech") # Switch to client variant
6. scaffold_preset({ preset_id: "client-x" })# Create new client preset
Creating New Presets
New presets only need to override tokens that differ from the parent:
// presets/client-newbrand/tokens.json
{
"colors": {
"accent": {
"primary": "#e11d48"
}
}
}
// presets/client-newbrand/manifest.json
{
"id": "client-newbrand",
"name": "New Brand",
"extends": "glassmorphic-base",
"version": "1.0.0",
"tags": ["custom"],
"components": [],
"layouts": []
}
All base tokens, components, and layouts are inherited automatically.
Correction Rules
The correction engine enforces these rules on every component:
- no-hardcoded-colors (error): All color values must use CSS custom properties
- no-hardcoded-spacing (warning): Spacing should use the token scale
- enforce-glass-surface (error): Surface elements must have
backdropFilter+ semi-transparent bg - no-hardcoded-font-family (error): Font families must use typography tokens
- use-animation-tokens (warning): Transitions must use animation tokens
- enforce-sidebar-components (error): Sidebar content must use NavGroup/NavItem
- enforce-settings-components (error): Settings UI must use OptionGroup/OptionRow
- a11y-img-alt (warning): Images must have alt attributes
- a11y-icon-button-label (warning): Icon buttons need aria-label
Monaco Integration
In your Monaco editor, call autocorrect_component on the save event:
editor.onDidSaveModel(async () => {
const code = editor.getValue();
const result = await mcpClient.callTool('autocorrect_component', {
code,
context: 'auto',
dry_run: false
});
if (result.corrected !== code) {
editor.setValue(result.corrected);
}
});
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.