Schema Awareness
When you connect a database to Grepp, the agent automatically learns your schema — your tables, columns, data types, and relationships. This means you can ask questions about your data in plain language, and the agent will write accurate SQL without you ever having to describe your data model.
Automatic schema discovery
As soon as a database integration is connected, Grepp scans it and builds an understanding of its structure. This includes:
- All available tables
- Column names and data types
- Nullability constraints
This schema is automatically included in the agent's context on every conversation turn, so it's always working with an up-to-date view of your data.
What this means for you
Because the agent already knows your schema, you can ask questions naturally:
- "Show me total revenue by month" — the agent finds the right table and date column
- "Which customers have placed more than 5 orders?" — the agent knows how orders and customers relate
- "What's the average order value this quarter?" — the agent uses your actual column names
You don't need to tell the agent what your tables are called or what columns exist. It already knows.
Supported databases
Schema awareness works automatically for:
- PostgreSQL
- MySQL
- SQL Server (MSSQL)
- ClickHouse
Large schemas
For databases with very large schemas, Grepp focuses the agent's attention on the most relevant tables for each query. If the agent needs more detail about a specific table, it will look it up. You can also guide it by mentioning a table or column name directly in your message.
Keeping schema up to date
If you add new tables or columns to your database, Grepp will pick them up automatically. Schema is refreshed whenever the agent queries your database, so your context stays current as your data model evolves.