DatI: A Semantic Database Gateway for Safe and Efficient Agent Access

Описание и применение

DatI предоставляет безопасный и эффективный шлюз для доступа к базам данных с помощью AI‑агентов, используя протокол MCP и семантическое моделирование.

Задача

Business teams struggled to adopt natural language querying for databases because NL2DSL limited ad‑hoc analysis and customers wanted integration into existing workflows rather than standalone data‑agent apps.

Результат

DatI enables safe, efficient, and integrated NL2SQL access for AI agents, reducing the need for custom code and providing enterprise‑grade credential management and semantic modeling.

Рабочий процесс

• Connect a database and auto‑read schemas
• Configure table/column descriptions, aliases, and enum values
• Define semantic subjects and business terms
• Set up preset tools (metadata search, schema inspection, controlled SQL) and custom parameterized SQL templates
• Publish the MCP service with access tokens
• Connect an agent (e.g., Google Antigravity) to the MCP endpoint and run semantic queries

Доказательства

publishing a database directly as a standard MCP service—paired with lightweight metadata management—provides a clean, plug‑and‑play NL2SQL component

Источник

DatI connects directly to your database and reads schemas automatically. Here, you can configure descriptions and aliases for tables, columns, and column values

Источник