mcp-schema-designer

Schema Designer

MCP server for database schema design and SQL generation. Design schemas from natural language, validate them for best practices, generate migrations, produce typed outputs (TypeScript/Zod/JSON Schema), and optimize queries -- all targeting PostgreSQL.

Pure logic -- no database connection needed. It generates SQL, it does not run it.

Pricing

Plan	Price	Features
Free trial	$0	3 calls total (shared across all tools), no credit card	npx -y mcp-schema-designer
Basic	$10/mo	design_schema, validate_schema, generate_types	Buy →
Pro	$20/mo	All Basic tools + migrate_schema, optimize_queries	Buy →

License keys are delivered to your email immediately after checkout. More info: aivp-mcp.vercel.app

Installation

Note: the npm package is mcp-schema-designer (the bare name schema-designer is a different, unrelated package).

No install step needed — run straight from npm:

npx -y mcp-schema-designer

Or install globally:

npm install -g mcp-schema-designer

Free trial: 3 calls total (shared across all tools), no credit card. Buy a license at the links in Pricing — your key is emailed instantly. Activate it with the LICENSE_KEY environment variable.

Usage

stdio (default)

npx -y mcp-schema-designer

SSE (HTTP)

npx -y mcp-schema-designer --sse
# Listening on http://localhost:3000

Set a custom port:

PORT=8080 npx -y mcp-schema-designer --sse

Claude Desktop Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "schema-designer": {
      "command": "npx",
      "args": ["-y", "mcp-schema-designer"],
      "env": { "LICENSE_KEY": "<your license key — omit for free trial>" }
    }
  }
}

Tools

design_schema

Design a database schema from a natural language description. Generates CREATE TABLE SQL (PostgreSQL), index recommendations, and a text-based ERD.

Input:

{
  "description": "An e-commerce platform with products, orders, and user accounts"
}

Or provide explicit table definitions to validate and enhance:

{
  "description": "Custom blog schema",
  "tables": [
    {
      "name": "posts",
      "columns": [
        { "name": "id", "type": "SERIAL", "primary": true },
        { "name": "title", "type": "VARCHAR(255)" },
        { "name": "author_id", "type": "INTEGER", "references": "users.id" }
      ]
    }
  ]
}

validate_schema

Validate SQL CREATE TABLE statements for best practices.

Checks for:

• Missing primary keys
• Unnamed constraints
• Missing indexes on foreign keys
• Data type issues (e.g., TIMESTAMP vs TIMESTAMPTZ, FLOAT for money)
• Naming convention violations (snake_case, plural table names)
• N+1 query risks based on relationships

Input:

{
  "sql": "CREATE TABLE User (id SERIAL PRIMARY KEY, firstName VARCHAR(255), orderTotal FLOAT);"
}

migrate_schema

Generate migration SQL by diffing two schemas.

Input:

{
  "from_sql": "CREATE TABLE users (id SERIAL PRIMARY KEY, name VARCHAR(255));",
  "to_sql": "CREATE TABLE users (id SERIAL PRIMARY KEY, name VARCHAR(255), email VARCHAR(255) NOT NULL UNIQUE);"
}

Output: UP migration (ALTER TABLE to add email column), DOWN migration (rollback), warnings for destructive changes.

generate_types

Generate typed code from SQL schema. Supports three output formats.

Input:

{
  "sql": "CREATE TABLE users (id SERIAL PRIMARY KEY, email VARCHAR(255) NOT NULL UNIQUE, name VARCHAR(255), created_at TIMESTAMPTZ DEFAULT CURRENT_TIMESTAMP);",
  "format": "typescript"
}

Formats:

Format	Output
typescript	TypeScript interfaces with Create variants
zod	Zod validation schemas with inferred types
json-schema	JSON Schema (draft 2020-12) definitions

optimize_queries

Analyze a SQL query against a schema and suggest performance improvements.

Input:

{
  "query": "SELECT * FROM orders WHERE user_id = 123 AND status = 'active' ORDER BY created_at DESC OFFSET 500",
  "schema_sql": "CREATE TABLE orders (id SERIAL PRIMARY KEY, user_id INTEGER REFERENCES users(id), status VARCHAR(50), created_at TIMESTAMPTZ DEFAULT CURRENT_TIMESTAMP);"
}

Identifies:

• Missing indexes (WHERE, JOIN, ORDER BY columns)
• Query rewrites (SELECT *, NOT IN, leading wildcards, large OFFSET)
• EXPLAIN plan interpretation hints

Development

npm install
npm run dev    # Watch mode
npm run build  # Production build
npm start      # Run stdio

License

MIT