PcVue SQL Chat - Industrial Intelligence

Table / View

Ask a question in plain English. The AI reads your question and translates it into a database query. The database executes it and the real results are returned to you — the AI never sees your actual data.

Understand

The AI reads your question and identifies what equipment, sensors, and time range you're asking about.

"Show me Mash Tun temperature from last week" → equipment: MT01T, range: 7d

Write Query

Using only the database structure (not your data), it writes a precise SQL query to fetch exactly what you need.

SELECT name, value, timestamp FROM variables WHERE name = 'MT01T' ...

Side-by-side comparison: the AI only sees the table schema (column names and types), while the real database with actual sensor values remains locked behind a shield. — What does the AI actually see?

Execute & Respond

The database runs the query. Real results are inserted into a response template — the AI formats, your data stays real.

Database → real rows → "The Mash Tun temperature was 65.2 °C at ..."

The AI only sees table structure while the actual database values remain protected — structure only, values never exposed. — Are the numbers invented?

Your data never reaches the AI

The AI sees your question, equipment names, and the database structure — just enough to write the right query. It never sees the actual database, sensor readings, values, or query results. All numbers in the response come directly from the database — nothing is invented.

Complete pipeline: User question flows to AI Translator (metadata only), then to SQL Query, through a privacy shield the AI cannot cross, to the Database (real data), and finally to a formatted Response. — What actually happens?

Zero Data Exposure

Production values, sensor readings, and operational data stay inside your database. The AI only works with metadata.

No Hallucinations

Every number in the response comes from an actual database query. The AI formats — it doesn't fabricate data.

Natural Language

Ask in plain English — no SQL knowledge needed. The system handles translation and execution automatically.

Welcome to SQL Chat. Ask questions about the brewery SCADA data and I'll generate SQL queries to find the answers.

LLM Context - Catalog & Schema

LLM Catalog Context (Step 1: Analyzer)

Compact Catalog (Step 2: SQL Generator)

LLM Schema Context (Step 2: SQL Generator)

Database Table Browser

How It Works

Your data never reaches the AI

Zero Data Exposure

No Hallucinations

Natural Language

Report Issue