SkematIA Documentation

What is SkematIA?

SkematIA is a Backend as a Service that lets you design, build, and query databases using natural language and AI. It runs on Firebird 5.0 and exposes your data through two interfaces: a Model Context Protocol (MCP) server for AI agents, and a REST API for any HTTP client.

You don't need database drivers, SQL knowledge, or direct Firebird access. Describe your data in plain language, and SkematIA creates the schema, tables, relationships, and constraints automatically. Then query your data conversationally or with raw SQL.

Quickstart (under 10 minutes)

1. Sign up at skematia.com and create a project.

2. Copy your API key (starts with sk_).

3. Choose your connection method: MCP for AI agents, or REST for any HTTP client.

4. Set header: Authorization: Bearer sk_your_api_key

5. Start with list_tables or get_schema to explore your database.

Connect via MCP (for AI Agents)

SkematIA exposes a Streamable HTTP MCP server. It's stateless: each request is independent, no session management needed.

Endpoint

https://skematia.com/mcp/mcp

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "skematia": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://skematia.com/mcp/mcp",
        "--header",
        "Authorization: Bearer sk_YOUR_API_KEY"
      ]
    }
  }
}

Claude Code (CLI)

claude mcp add skematia --transport http https://skematia.com/mcp/mcp

Python MCP SDK

from mcp.client.streamable_http import streamablehttp_client
from mcp import ClientSession

async with streamablehttp_client(
    url="https://skematia.com/mcp/mcp",
    headers={"Authorization": "Bearer sk_YOUR_API_KEY"},
) as (read, write, _):
    async with ClientSession(read, write) as session:
        await session.initialize()
        tools = await session.list_tools()
        result = await session.call_tool("query", {
            "project_id": "your-project-id",
            "question": "how many tables exist?",
        })

Raw HTTP (curl)

curl -X POST https://skematia.com/mcp/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk_YOUR_API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/list",
    "params": {}
  }'

REST API

All REST endpoints require the /api/ prefix. Without it, requests hit the frontend and return 405. No trailing slashes.

SQL dialect is Firebird 5.0: column aliases require AS, table names are UPPERCASE, use ROWS N instead of LIMIT N.

Schema & Explorer

Data Operations

AI Studio

Projects & API Keys

Example (curl)

curl -H "Authorization: Bearer sk_YOUR_API_KEY" \
  https://skematia.com/api/explorer/{project_id}/schema

Example (Python)

import httpx
r = httpx.get(
    "https://skematia.com/api/explorer/{project_id}/schema",
    headers={"Authorization": "Bearer sk_YOUR_API_KEY"}
)
schema = r.json()

MCP Tools (15)

Security

SkematIA implements a Zero-Trust security model with three pillars:

Granular Permissions

Each API key has a permission level that controls allowed operations:

Audit Trail

Every operation is logged with user, project, action, SQL statement, duration, and result. The audit log is immutable and stored in a central database.

Rate Limiting

Sliding window rate limiting per API key. Response headers include X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset.

Resources

SkematIA Dashboard - Sign up, create projects, manage API keys.

Skematia Protocol v1.0 - Open specification for AI agents and IDEs to interact with databases. Apache 2.0 licensed.

Agent Card (A2A) - Machine-readable service descriptor for AI agents.

MCP Connection Guide - Step-by-step connection instructions.

Contact: soporte@skematia.com