AI assistants use Model Context Protocol tools to manage your database without writing code.

​
Available Tools

ToolPurpose
create-tableDesign and create tables
modify-tableAdd, drop, or rename columns
delete-tableRemove tables permanently
get-table-schemaView table structure
get-db-apiGet database API documentation

​
How AI Creates Tables

When you say “I need a blog”, the AI:
  1. Calls create-table with appropriate schema
  2. Adds relationships to _user table
  3. Creates indexes for performance
  4. Returns the API endpoints ready to use
// You say: "Create a products table"
// AI executes:
mcp.create_table({
  name: "products",
  columns: [
    { name: "id", type: "UUID", primary: true },
    { name: "name", type: "STRING" },
    { name: "price", type: "FLOAT" },
    { name: "stock", type: "INTEGER", default: 0 }
  ]
})

​
AI Understanding

The AI knows:
  • _user table is read-only via database
  • Foreign keys create relationships automatically
  • POST requires array syntax [{...}]
  • Which columns need indexes
  • How to handle timestamps and defaults

​
Workflow

  1. First Time: AI calls get-instructions to understand your backend
  2. Exploration: Uses get-table-schema to see existing tables
  3. Creation: Designs optimal schema with create-table
  4. Implementation: Uses API endpoints from get-db-api
  5. Debugging: Calls debug-backend if something fails
The AI handles all the complexity. You just describe what you need.