Antigravity MCP v1.6 E-E-A-T Verified TypeScript & MIT

HumanCraft UI — Antigravity MCP Server for Human Web Design

The premier Model Context Protocol (MCP) server designed exclusively for Google Antigravity (AGY), Claude Desktop, and Cursor. Eliminates "AI Slop" in website layouts and copy, enforces authentic E-E-A-T entity reconciliation, and crafts search-resilient digital experiences.

Protocol
Model Context Protocol (MCP)
Supported Clients
Google Antigravity, Claude, Cursor
License & Privacy
MIT / Zero Data Collection
Maintainer
AI Build Infra Engineering

Eliminating Algorithmic "AI Slop" & Penalties

When AI models generate landing pages without heuristics, they fall into generic templates that fail Google's Information Gain ranking models.

The AI Slop Vulnerability

Triggers Google SpamBrain Demotions
  • Generic Heading Tropes: Headings like "What is X?" and "Key Benefits" offer zero information gain.
  • Hallucinated Metrics & Personas: Fake statistics ("boosted efficiency by 30%") trigger manual quality actions.
  • Binary Checkmark Tables: Useless Yes/No comparison tables that fail to communicate genuine architectural trade-offs.
  • Visual Homogeneity: Monotonous purple/pink gradients, stiff 3-card grids, and dead buttons without tactile feedback.

The HumanCraft Defense

Google Information Gain Compliant
  • Outcome-Driven Intent Headings: Enforces empirical titles with concrete numbers and specific technical tipping points.
  • Strict Anti-Default Intake: Rejects synthetic placeholders; mandates real Day 0 vs Day 90 practitioner data.
  • Empirical Trade-off Matrices: Generates multi-parameter matrices with unit economics and semantic ItemList schema.
  • Tactile Human Archetypes: Curated design systems (Editorial Tech, Dark Craft, Warm Humanist) with rich physics.

The 6 Core HumanCraft MCP Tools

A specialized suite of linters, graph reconcilers, and design generators operating directly inside your AI coding assistant.

📋

harvest_client_intake

Extracts real practitioner experiences, telemetry screenshots, and baseline vs audited metrics. Strictly forbids default placeholders.

Strict Anti-Defaulting
🎯

validate_heading_intent

Lints headings against generic AI tropes. Scores and enforces outcome-driven H1-H3s aligned with US Patent 2022/0277032 A1.

Information Gain
📊

build_comparison_matrix

Outlaws binary Yes/No tables. Generates multi-parameter comparison matrices with unit economics and Schema.org ItemList markup.

Semantic JSON-LD
🛡️

audit_eeat_entity_graph

Reconciles author and publisher JSON-LD against Google's Knowledge Graph standards. Verifies authoritative external sameAs URIs.

Knowledge Graph
🔍

lint_design_anti_patterns

Scans HTML and Tailwind snippets for purple/cyan gradients, rigid 3-card rows, missing button squeeze physics, and AI buzzwords.

Anti-Slop Linter
🎨

get_human_archetype

Returns curated design systems with tinted neutrals, tactile physics, and typography: Editorial Tech, Tactile Dark Craft, Warm Humanist, Neo-Brutalist.

Design Archetypes

Add HumanCraft to Your Assistant in 60 Seconds

Compatible with Google Antigravity, Claude Desktop, Cursor, and any MCP-compliant AI assistant.

mcp_config.json
{
  "mcpServers": {
    "humancraft": {
      "command": "node",
      "args": [
        "C:/path/to/Humancraft-UI/dist/index.js"
      ]
    }
  }
}
Clone directly from GitHub: git clone https://github.com/AI-BuildInfra/Humancraft-UI.git
View Repository
Looking for Core Web Vitals & Search Intent Diagnostics?
Explore SEO-IntentRank MCP

Need Custom AI Tooling or High-Performance Web Infrastructure?

From custom Model Context Protocol servers to enterprise Next.js platforms, AI Build Infra architects solutions engineered for real business outcomes.

Chat with us!
Get Free Quote