SEMANTIX docs
Text-to-SQL is a solved problem. Trustworthy numbers are not.
SEMANTIX is the context layer that anchors every answer in governed, centralized context: certified metrics, living business knowledge, and readable execution.
Start here
Section titled “Start here”- Quickstart — three steps to a working tenant
- What is SEMANTIX — the concepts and vocabulary
- Analyst — the day-to-day experience for querying data
- Admin — providers, connectors, scopes, and governance
The one-page version
Section titled “The one-page version”SEMANTIX connects to your databases as read-only data sources, catalogues what it finds in a Registry (entities and metrics with real descriptions), groups that context into Context Scopes shaped like your org, and lets people ask questions in plain language. Answers are grounded in that governed context rather than generated from raw schema, so the numbers are the ones your team agreed on.
It is also a semantic MCP server — the same governed context is available to any AI agent that speaks the Model Context Protocol.
What it connects to
Section titled “What it connects to”Any warehouse with a SQL interface, plus anything that speaks MCP. Full field detail is in the data source reference.
PostgreSQL 
MySQL 
SQL Server 
Snowflake 
Databricks 
BigQuery 
Redshift 
Oracle 
ClickHouse 
Trino Over MCP: Confluence, Notion, GitHub, Slack, Jira, Google Drive, Linear, Salesforce.
A five-minute orientation
Section titled “A five-minute orientation”| If you want to… | Go to |
|---|---|
| Get a tenant running | Quickstart |
| Understand the terms | What is SEMANTIX |
| Know why not to just use a coding agent | vs coding agents |
| See what a new tenant starts with | New tenant reference |
| Fix an error | Troubleshooting |