SkematIA bronze

AI Schema Designer

Design a complete Firebird 5.0 database schema from a natural language description. The AI generates tables, fields, relationships, constraints, and executes everything automatically.

• Create a schema for an e-commerce store with products, orders, and customers
• Design a project management database with tasks, teams, and deadlines

AI Schema Modifier

Modify an existing database schema using natural language instructions. Generates ALTER TABLE, CREATE TABLE, or other DDL statements and executes them.

• Add an email field to the users table
• Create a products table with price and inventory fields

Natural Language Query

Query your database using natural language. The AI converts your question to Firebird SQL and returns the results.

• How many products cost more than 100?
• Show me the top 5 customers by total orders

Execute SQL

Execute raw SQL against your Firebird database. Allowed operations depend on the API key's permission level: read_only allows SELECT only, read_write adds DML, full allows DDL. All executions are audit-logged.

• SELECT * FROM "PRODUCTS" ROWS 10
• INSERT INTO "USERS" ("NAME", "EMAIL") VALUES ('Juan', 'juan@test.com')

Get Schema

Retrieve the complete schema of a project including tables, fields, primary keys, foreign keys, and relationships.

List Tables

List all tables in a project's database with field counts.

Insert Data

Insert one or more records into a table using column-value pairs.

• Insert a new product with name 'Widget' and price 29.99

Update Data

Update a single record by its primary key.

Delete Data

Delete a single record by its primary key.

Bulk Update

Update multiple records matching a WHERE condition.

List Projects

List projects accessible with the current API key.

Create Project

Get instructions for creating a new SkematIA project with a Firebird database.

Embedded Project Management

Create the 8 _CTRL_ tables in your project's database so agents can track tasks, sessions, findings, decisions, knowledge and metrics directly in the same DB they work on — persistent shared memory across sessions and across agents. Idempotent; not available on the FREE plan; requires full permissions. Table names start with _ so they need double quotes in SQL (SELECT * FROM "_CTRL_TAREAS").

• Set up control tables so I can track my progress in this project
• Give this project persistent agent memory

List Backups

List available backups for a project with size, date, and type (scheduled, pre-restore, manual).

Schema Advisor

Analyze the project's database schema and return improvement suggestions: missing indexes on FK columns, fields that should be NOT NULL or UNIQUE, tables without primary key, and BLOB TEXT used for short fields. Each suggestion includes severity, description, and a ready-to-execute SQL fix.

• Analyze my schema for potential improvements
• Check if any foreign keys are missing indexes

SchemaAgent — Semantic Schema Analysis

Analyze the schema using REAL data: samples up to 50 rows per table and infers each field's semantics from name + type + observed values. Detects money stored as FLOAT (suggests NUMERIC(15,2)), dates/numbers stored as text, consistent formats (email, RFC, postal code → CHECK constraints), implicit catalogs, repeated fields across tables (→ DOMAIN), and unindexed filter fields. Every suggestion includes a ready-to-execute sql_fix. Read-only and domain-agnostic.

• Analyze my schema semantically and tell me what to improve
• Are there fields storing money as float or dates as text?

Schema Version History

Retrieve the DDL change history for a project's database. Every CREATE, ALTER, DROP, and RECREATE operation is automatically recorded with version number, timestamp, source (ai_studio, explorer, mcp, api), and executing agent. Supports filtering by limit, source, and operation type.

• Show me the last 10 schema changes
• What DDL changes were made by the AI studio?

Inspect Objects

Ver triggers, procedures y generators CON su codigo fuente/definicion (lo que get_schema no incluye).