(opens in a new tab)

Azure Data Studio on Mac: Status and Alternatives

September 2, 2026 by Chat2DB Team

If you work with SQL Server on a Mac, Azure Data Studio was the answer for most of a decade. SSMS has never had a macOS build and never will, so ADS was the one Microsoft-made client that ran natively on a Mac. That ended on 28 February 2026, when Azure Data Studio reached end of support.

This article covers what still works, what to do about it, and how to connect a Mac to SQL Server without a Windows virtual machine.

What still works on a Mac

An installed copy of Azure Data Studio launches and connects as before. Nothing was disabled remotely. But there are no more releases, which means no security patches to the Electron runtime it is built on, no compatibility fixes for newer SQL Server versions, and no support if something breaks. On a managed corporate Mac, that combination usually fails a security review within a release cycle or two.

There is also a practical Apple Silicon wrinkle. Later ADS builds were universal binaries, so they run natively on M-series chips, but the version you have installed may predate that and be running under Rosetta 2. Check with:

# Look at the architectures the binary actually contains
lipo -archs "/Applications/Azure Data Studio.app/Contents/MacOS/azuredatastudio"
 
# And whether the running process is translated (1 = under Rosetta)
sysctl -n sysctl.proc_translated

If that shows x86_64 alone, you are running an Intel build through translation on a machine that will never get a native update.

The command line still works fine

Before choosing a GUI, it is worth knowing that Microsoft's command line tools are supported on macOS and are the most reliable fallback:

# Install the driver and tools (Homebrew)
brew tap microsoft/mssql-release https://github.com/Microsoft/homebrew-mssql-release
brew update
brew install msodbcsql18 mssql-tools18
 
# Connect. -C trusts the server certificate; drop it once you have a real one.
sqlcmd -S sql01.internal,1433 -U reporting -P "$SQLPASS" -d appdb -C
 
# One-off query, tab separated
sqlcmd -S sql01.internal -U reporting -P "$SQLPASS" -d appdb -C \
  -Q "SELECT name, database_id, state_desc FROM sys.databases ORDER BY name" -s $'\t' -W

There is also go-sqlcmd, the newer Go implementation, which handles Entra ID authentication more cleanly:

brew install sqlcmd
sqlcmd config add-endpoint --name sql01 --address sql01.internal --port 1433
sqlcmd query "SELECT @@VERSION"

And running SQL Server itself locally on Apple Silicon is straightforward with the Azure SQL Edge or SQL Server container images:

docker run --platform linux/amd64 -d --name mssql \
  -e "ACCEPT_EULA=Y" -e "MSSQL_SA_PASSWORD=Str0ng!Passw0rd" \
  -p 1433:1433 mcr.microsoft.com/mssql/server:2022-latest

On M-series Macs the SQL Server image runs under emulation, which is slow but adequate for development. Azure SQL Edge has native arm64 images if you need the performance and can live with its smaller feature set.

Choosing a GUI replacement

Four options genuinely work on macOS, including on Apple Silicon.

Chat2DB. The closest thing to a drop-in replacement for how most people used ADS. It connects to SQL Server over TDS — named instances, Azure SQL Database and Azure SQL Managed Instance included — plus PostgreSQL, MySQL, Oracle, MongoDB, Redis and 20+ other engines from a single window. Its AI writes and explains T-SQL against your actual schema, which covers a lot of what ADS notebooks were used for: understanding an inherited stored procedure, drafting a query against tables you did not design. Available from chat2db.ai/download (opens in a new tab), with a browser version at app.chat2db.ai (opens in a new tab) if you want to check a server without installing anything.

DBeaver Community. Free, open source, and supports more database engines than anything else. The interface is Eclipse-based and feels dated next to a native Mac application, and it wants a fair amount of memory, but for breadth of coverage nothing beats it.

DataGrip. JetBrains' database IDE. Excellent SQL editing, refactoring and version control integration. Worth the subscription if you already pay for JetBrains tools; hard to justify on its own if you only need a query window.

TablePlus. The most Mac-native of the four: fast to launch, small memory footprint, a genuinely good results grid. Fewer administration features, and the free tier limits open tabs.

All four connect to SQL Server without a virtual machine, and all four run natively on Apple Silicon.

Connection settings that trip Mac users up

Three issues account for most first-connection failures from macOS.

Certificate validation. ODBC Driver 18 defaults to Encrypt=yes, so a server with a self-signed certificate is rejected. The quick unblock is trusting the certificate; the correct fix is installing one the Mac's keychain trusts:

-- Confirm what the current session negotiated
SELECT session_id, net_transport, encrypt_option, auth_scheme, client_net_address
FROM sys.dm_exec_connections
WHERE session_id = @@SPID;

If encrypt_option reads FALSE, the connection is in the clear and worth fixing before it becomes an audit finding.

Named instances. The SQL Browser service resolves SERVER\INSTANCE names over UDP 1434, which is frequently blocked across a VPN. Use the explicit port instead:

-- Run on the server to find the TCP port an instance is listening on
SELECT DISTINCT
       local_tcp_port,
       SERVERPROPERTY('InstanceName') AS instance_name
FROM sys.dm_exec_connections
WHERE local_tcp_port IS NOT NULL;

Then connect to sql01.internal,1435 rather than sql01\SQLPROD.

Windows authentication. Integrated auth from a Mac needs Kerberos, which means the Mac must be joined to or have a ticket from the domain:

kinit user@CORP.EXAMPLE.COM
klist

With a valid ticket, clients that support it can use Authentication=ActiveDirectoryIntegrated. Where that is impractical, Microsoft Entra ID authentication against Azure SQL is generally easier, and a SQL login is the pragmatic fallback for on-premises servers.

A short migration path

1. Export your ADS connection profiles from ~/Library/Application Support/azuredatastudio/User/settings.json — the datasource.connections key. Passwords live in the Keychain, not the file, so you will re-enter them.
2. Convert any SQL notebooks to .sql scripts or Jupyter notebooks while ADS is still installed and can open them.
3. Install one replacement client and rebuild two or three connections. Verify: object explorer, a query with results, an export, and a plan.
4. Install mssql-tools18 regardless of which GUI you pick. Having sqlcmd on the machine has saved more incidents than any GUI feature.
5. Remove Azure Data Studio once the replacement covers your daily work, and take it off the software allow-list.

What you will miss, and what you will not

It helps to be specific about the gap, because a lot of the anxiety around leaving Azure Data Studio turns out to be about features people rarely used.

You will not miss much for query work. Every client listed above gives you IntelliSense, a results grid, multi-statement batches, result export and schema browsing. Most give you better grid editing than ADS had.

The two real losses are notebooks and the server dashboards. Notebooks are recoverable — the file format is standard .ipynb, so the content moves to Jupyter or to plain .sql scripts with the narrative kept as comments. The dashboards are not, but they were mostly thin wrappers over DMV queries that you can keep as saved scripts instead:

-- Dashboard replacement: database sizes, biggest first
SELECT DB_NAME(database_id) AS database_name,
       CAST(SUM(size) * 8.0 / 1024 AS decimal(10,1)) AS size_mb,
       type_desc
FROM sys.master_files
GROUP BY database_id, type_desc
ORDER BY size_mb DESC;
 
-- Dashboard replacement: what is running right now
SELECT r.session_id, r.status, r.command, r.wait_type,
       r.total_elapsed_time / 1000.0 AS elapsed_seconds,
       s.login_name, s.program_name
FROM sys.dm_exec_requests AS r
JOIN sys.dm_exec_sessions AS s ON s.session_id = r.session_id
WHERE r.session_id <> @@SPID AND s.is_user_process = 1
ORDER BY r.total_elapsed_time DESC;

Keep those in a Git repository rather than in a client-specific dashboard format, and the next tool migration costs nothing.

What about Parallels or a Windows VM?

You can run SSMS in a Windows VM on a Mac — under Parallels on Apple Silicon this means Windows on ARM, which runs SSMS through x64 emulation. It works, and for a DBA who needs Maintenance Plans, Always On wizards or the Replication Monitor there is no alternative, because those genuinely exist nowhere else.

For everyday query work it is a poor trade: several gigabytes of RAM, a Windows licence, a second desktop to context-switch into, and clipboard friction. Use a native client for the 95% of work that does not need a wizard, and keep a VM or a jump box for the 5% that does.

Summary

Azure Data Studio was the Mac answer for SQL Server and it stopped being supported on 28 February 2026. Installed copies still run, but with no security patches they are hard to keep on a managed machine. SSMS is not and never has been an option on macOS. The replacements that work natively, Apple Silicon included, are Chat2DB, DBeaver, DataGrip and TablePlus — with Chat2DB the closest match if you also touch PostgreSQL, MySQL or MongoDB alongside SQL Server. Install mssql-tools18 for the command line either way, export your connection profiles before uninstalling, and keep a Windows VM only for the administration wizards that genuinely exist nowhere else.