Free & open source · Public Preview

See how your SQL database handles a workload.

Configure queries, run concurrent users, and watch performance live with sqlsim Studio. Go from a single query to a repeatable load test in a local browser app.

v1.0.0 · Windows x64 · Includes Studio and both CLIs.
Requires ODBC Driver 18 and access to a SQL database. No .NET install needed. Installation guide

See sqlsim Studio in action
sqlsim Studio workload configuration and live monitoring interface
Configure a workload, watch live monitoring, then review and compare results. Read the Studio walkthrough →

Four ways to work

Start visually, automate from the command line, compare drivers, or describe the experiment in plain language.

1

sqlsim Studio

Configure, monitor, replay, and compare workloads in a self-contained local browser app.

Explore Studio →
2

sqlsim

Native C++ and ODBC for low-overhead scripts, concurrent load tests, and replay.

Read the CLI guide →
3

sqlpysim

A Python workflow powered by Microsoft's mssql-python driver for automation and comparison.

See Python examples →
4

Workload Builder

Use GitHub Copilot and other agents to design and compare experiments from plain-language requests.

Use with agents →

One toolkit, every SQL Server target

Use the same workload patterns across local development, Azure, and Microsoft Fabric.

✓SQL Server 2019, 2022, and 2025
✓Azure SQL Database and Managed Instance
✓SQL database in Microsoft Fabric
✓Windows and Ubuntu Public Preview paths

Run your first workload

Start with Studio on Windows. No hosted service, no .NET install, and no telemetry sent elsewhere. Use a development or test database you have permission to load test.

1
Download and extract →Install Microsoft ODBC Driver 18. Download the Windows bundle, unblock the trusted ZIP in Properties, then extract it. Keep the folder structure intact.
2
Open Studio and connectDouble-click sqlsimstudio\sqlsimstudio.exe. In your browser, choose your SQL server, database, and authentication method.
3
Run, watch, and compare →Enter a query, start with a small thread and iteration count, and run it. Watch live progress and review results before increasing the load.