SQLDesk
Query 35+ data sources, build visualizations and dashboards on what comes back, and put the result on a wall or in somebody's inbox. Self-hosted, Apache 2.0, no per-seat pricing.
What it does
You write SQL against a warehouse. SQLDesk runs it, keeps the result, draws it, and puts it on a dashboard that refreshes itself. That is the whole loop, and everything else here is a detail of it.
- Queries are SQL you save, with parameters if you want them. They can run on a schedule so a dashboard is ready before anyone opens it.
- Visualizations belong to a query. One query can have many — a table, a chart and a number are three views of the same rows, not three queries.
- Dashboards arrange visualizations from any number of queries, with filters that drive them all.
- Alerts watch a query's result and tell you when it crosses a line.
What it is not
SQLDesk does not move or store your data. It holds query results for as long as you configure, and its own metadata — users, queries, dashboards. The warehouse stays where it is and remains the source of truth. There is no ETL here and no modelling layer you have to fill in before anything works.
Where to go next
- Concepts
- The words this documentation uses, and what each one means precisely.
- Architecture
- What the moving parts are — Postgres, Redis, workers — and why each one is there.
- Deploying
- Docker Compose, Kubernetes, and what has to be configured before it will start.