---
title: "ai-ui-cleaner"
description: "Create, review, and repair authored interfaces using retrieved code and visual references."
canonical: https://agentpluginsdirectory.com/plugins/ai-ui-cleaner
last-updated: 2026-10-21
---

# ai-ui-cleaner
Create, review, and repair authored interfaces using retrieved code and visual references.
- Slug: ai-ui-cleaner
- Publisher: Mustafa Hameed
- Repository: https://github.com/MustafaHam123/ai-ui-cleaner
- Manifest: plugin.json
- Version: 0.1.3
- License: MIT
- Category (editorial): other
- Skills: 1 (ai-ui-cleaner)
- MCP servers: 1 (ai_ui_cleaner)
- Stars: 0
- Repository created: 2026-10-03
- Repository last pushed: 2026-10-08
- Publisher type: User
- Listing: https://agentpluginsdirectory.com/plugins/ai-ui-cleaner
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What ai-ui-cleaner does, in the publisher's words

AI UI Cleaner bundles a design skill with a public reference-search MCP for Codex (desktop and terminal) and Claude Code (Code tab and terminal). Install the plugin once to get both. No separate skill download, corpus download, Cloudflare account, Node runtime or MCP API key is needed. Your chosen host still needs its normal account/login and internet access.

- Resolve the target: mobile app, mobile web, desktop app/web, tablet or responsive website. Ask when unclear.
- Resolve the vibe from the brief or supplied design; ask when missing, rather than infer it from the product category.
- Search the RAG library for that surface and visual structure, not simply the product prompt.
- Inspect 3 to 5 matching screenshots and choose one primary frame.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/MustafaHam123/ai-ui-cleaner/HEAD/README.md

## MCP servers

- ai_ui_cleaner: transport: http; url: https://ai-ui-cleaner.ai-ui-cleaner.workers.dev/mcp

Read from the plugin's own mcp.json. Environment variable names only, never values.