Log inSubmit server
Aesthete Agent logo

Aesthete Agent

Validates and exports frontend color systems with palette matching, OKLCH harmony, contrast checks, simulations, and design-token outputs.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Aesthete Agent

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "docsor1212-aesthete-agent": {
      "command": "npx",
      "args": [
        "-y",
        "clawhub"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

The aesthete-agent MCP server connects AI coding agents to a local frontend color-system engine through callable MCP tools. It ranks 118 curated palettes, generates OKLCH harmonies, checks WCAG 2.2 and optional APCA contrast, simulates color-blindness, detects AI-slop signatures, and exports CSS, Tailwind, W3C, context, or prompt formats. Reach for it when an agent needs repeatable palette selection and measurable validation rather than unconstrained color suggestions.

Use cases

  • Rank palettes for a frontend design brief
  • Generate accessible OKLCH color harmonies
  • Check WCAG and APCA contrast before export
  • Find color-blind collisions and AI-slop patterns
  • Export validated colors as CSS or design tokens

Key features

  • Evidence-based palette matching
  • OKLCH harmony generation
  • WCAG 2.2 and APCA contrast checks
  • Color-blindness simulation
  • AI-slop and gradient detection
  • CSS, Tailwind, and W3C token export

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 3d ago

Recent health check succeeded.

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 3d ago via OSV.dev · clawhub (PyPI)

How Aesthete Agent works

What aesthete-agent MCP server does

The aesthete-agent MCP server gives MCP-compatible agents structured operations for planning, testing, and exporting frontend color systems. Its engine combines an evidence library of 118 curated palettes with measurable checks instead of treating palette choice as an unvalidated design suggestion.

The server can rank palettes against a short design brief, generate color relationships in OKLCH, assign or inspect semantic roles, and evaluate a palette with a five-dimension scorecard. It also produces pairwise and full-matrix contrast results using WCAG calculations, with APCA available for matrix checks. Color-blindness simulation and categorical color generation help identify collisions that ordinary contrast ratios do not cover.

How it works

MCP tools communicate with the bundled engine over JSON-RPC on standard input and output. The package includes the engine data, so the server can run as a standalone wheel rather than depending on a separate checkout.

Validation is organized around several gates. These include palette scoring, contrast boundaries, color-blind collision checks, AI-slop signature detection, and an HTML delivery linter. A gradient is treated as a P0 failure by the quality gates. The palette_report tool combines the checks into a single PASS or FAIL result. Preference decisions can also be stored in profile.json, allowing later matching, scoring, and linting to down-rank rejected patterns.

The broader method follows six stages: establish a design direction, check the register, select a source palette, derive roles, validate the result, and review it through five lenses. The visualizer can render ten frontend scene types, including light and dark variants, so an agent or user can inspect the palette in context.

Setup and configuration

Run the MCP package with uvx aesthete-agent-mcp. This uses the PyPI distribution and does not require credentials. A pure-stdlib stdio fallback is available with the --stdio option when FastMCP is not installed. Smithery also provides an installation route through its CLI:

Terminal
npx -y @smithery/cli install docsor1212/aesthete-agent-mcp --client claude

The project is also distributed as an agent skill, but that is separate from the MCP server. Installing the skill uses npx skills add docsor1212/aesthete-agent and exposes the command-line engine rather than only the MCP wrapper.

Tools and capabilities

The aesthete-agent MCP server exposes tools for:

  • Matching palettes from the evidence library to a brief.
  • Generating OKLCH harmonies and checking palette scores.
  • Testing individual contrast pairs or complete contrast matrices with WCAG and APCA options.
  • Simulating color-blindness and producing color-blind-safe semantic colors.
  • Detecting AI-slop signatures.
  • Rendering ten UI scene types with optional variants.
  • Selecting named theme packs.
  • Exporting CSS, Tailwind, W3C Design Tokens, context, or image-generation prompt data.

Named theme packs contain colors, fonts, voice, intended use, and provenance. The underlying engine also supports additional operations such as ramps, mutations, role derivation, reports, visual critiques, health checks, and image extraction, although the MCP table specifically documents the tools above.

Limitations and notes

The server validates color-system properties; it does not replace contextual design review. The evidence library uses crowd signals and objective filters, while final in-context acceptance remains a user decision. Pillow is optional for image extraction, and FastMCP is optional because the stdio fallback can run without it.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Aesthete Agent

aesthete-agent is an MCP server that connects AI coding agents to a local frontend color-system engine. Its main tools match palettes, generate OKLCH harmonies, score and report quality, check WCAG or APCA contrast, simulate color-blindness, detect AI-slop signatures, visualize scenes, select themes, and export CSS or design tokens.

Install it with `uvx aesthete-agent-mcp`. The package runs from PyPI without API keys, and Smithery provides an alternative installation through its CLI. A stdio fallback is available with `--stdio` when FastMCP is not installed.

Configure Cursor to launch the server using the `uvx aesthete-agent-mcp` command.

Yes, aesthete-agent can export W3C Design Tokens JSON through its palette export capability. The same export tool also supports CSS, Tailwind, context, and prompt formats.

Related MCP servers

All in Architecture & Design All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewAesthete Agent AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/docsor1212-aesthete-agent?style=directory)](https://allmcps.com/mcp/docsor1212-aesthete-agent)
HTML Embed
<a href="https://allmcps.com/mcp/docsor1212-aesthete-agent"><img src="https://allmcps.com/api/badge/docsor1212-aesthete-agent?style=directory" alt="Aesthete Agent on AllMCPs" /></a>