⚒️ QuerySculptor

🚀 GraphQL for AI Agents

A server application that helps AI agents interact with GraphQL APIs through structured tools instead of raw query generation.

QuerySculptor provides 26 tools via the Model Context Protocol (MCP) that allow agents to build, validate, and execute GraphQL queries step-by-step rather than generating complete query strings.

🎪 Demo and Setup

🌍 Public Demo

MCP Endpoint: https://agent-query-builder-toolbox.vercel.app/mcp

The demo is pre-configured to use the Pokemon API for testing Query Sculptorfeatures.

🎮 Demo: Pokemon API Examples

Our demo uses the Pokemon API - here's what you can ask:

Discover Pokemon:

• "Show me the first 10 Pokemon with their types and sprites"
• "Find all Pokemon that are both Fire and Flying type"
• "What are the stats for Charizard?"

Explore Abilities & Moves:

• "List all Pokemon abilities and their effects"
• "What moves can Pikachu learn?"
• "Show me all Electric-type moves with their power and accuracy"

Regional & Species Data:

• "List all Pokemon from the Kanto region"
• "Show me Pokemon species with their evolution chains"
• "What berries are available and what do they do?"

Complex Queries:

• "Find Pokemon with abilities that boost attack in a pinch"
• "Show me all legendary Pokemon with their types and base stats"
• "List Pokemon that can learn both Water and Ice moves"

Add to your Cursor mcp.json:

{
  "mcpServers": {
    "graphql-query-builder-demo": {
      "url": "https://agent-query-builder-toolbox.vercel.app/mcp"
    }
  }
}

🏃‍♂️ Local Setup

# Clone and setup
git clone https://github.com/flinstonedev/querysculptor.git
cd querysculptor
pnpm install

# Configure (copy example.env to .env and customize)
cp example.env .env

# Launch
vercel dev
# 🚀 Your MCP server is live at http://localhost:3000/mcp