Skip to content
| Marketplace
Sign in
Visual Studio Code>Data Science>AddisDBNew to Visual Studio Code? Get it now.
AddisDB

AddisDB

Cenova Spark

|
1 install
| (0) | Free
Browse your databases and run SQL from VS Code — powered by the AddisDB CLI. Shares the connections you set up in the AddisDB desktop app.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

AddisDB for VS Code

Browse your databases, run SQL, and keep production safe — without leaving your editor.

Powered by the AddisDB CLI, sharing the exact connections and encrypted secret store you set up in the AddisDB desktop app. Add a connection in one, and it's there in the other.

Connects to 60 engines and file formats — PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, SQLite, MongoDB, Redis, Cassandra, Snowflake, BigQuery, Databricks, ClickHouse, DuckDB, Elasticsearch, Neo4j, Trino, TimescaleDB, pgvector, PostGIS, and CSV / Parquet / Excel files among them.


Requirements

  • A connection. Connections are created in the AddisDB desktop app or the addisdb CLI — this extension reads and uses them, it doesn't create them.
  • macOS, Linux, or Windows. macOS covers both Apple Silicon and Intel.
  • Internet access on first run, so the extension can fetch the CLI.

Zero setup

On first activation the extension downloads the small addisdb CLI, installs it to ~/.addisdb/bin, and puts it on your PATH — so addisdb also works in any terminal for the full-screen terminal UI. It keeps itself up to date; there's nothing to manage. Set addisdb.autoProvision to false if you'd rather install it yourself.


Safety, honestly

The reason this exists. Sessions open read-only. Writing is opt-in: a command that needs to write asks for a writable session, and every statement is still classified by the engine's safety layer, so a connection you've tagged read-only refuses the write at the source rather than in the UI. Result sets are auto-limited (addisdb.maxRows), so a stray SELECT * on a billion rows won't hang your editor.

That does not mean writes are impossible — the extension ships real schema and data operations. What it means is that nothing destructive happens quietly:

  • What does this statement do? classifies the statement under your cursor without touching the database — the same classification the engine uses to allow or refuse it. Ask before you run.
  • Test SQL in a sandbox provisions a throwaway copy (schema + data) and runs your SQL against that. The real database is never touched.
  • Estimate query cost and count rows answer "how bad is this?" first.
  • Drop table shows the row estimate and every dependent object, then offers to take a restorable backup before it proceeds.

AI that doesn't leak your data

AI features are bring-your-own-key and off until you set one (Set AI API Key). What a model can see is governed per connection:

  • AI Privacy Policy for This Database tells you exactly how much would be shared — "142 of 380 tables and 900 of 4,100 columns would be shared with a model" — and lets you narrow it.
  • Scan for sensitive columns finds what looks like regulated data and pre-selects it for locking, so accepting the safe default is one click.
  • Scan for Regulated Data (PII/PHI/PCI) audits a whole connection and reports what it classified, with a confidence score per column.
  • Generate SQL with AI (⌘I / Ctrl+I) writes the query from a prompt and tells you whether what it produced is read-only.

Everything else

  • Connections view — every connection in your workspace, grouped by project. Expand to browse schemas, tables, and columns; click a table to preview rows.
  • Run SQL — the whole file, the selection, or the statement under your cursor (⌘↩ / Ctrl+Enter, the ▷ Run CodeLens, or the editor toolbar). Multi-statement scripts return every result set.
  • Results grid — sortable and themed, with row count, timing, and copy-as-CSV.
  • Live monitoring — live metrics, and live activity where you can cancel or kill a running query.
  • Migrations — check migration status, review pending migrations with the safety layer's verdict on each, and generate new ones.
  • Mock data — generate realistic rows for a table, tracked in a ledger so you can clear exactly what you generated later, even from another machine.
  • Local servers — start and stop local database servers from the editor.

Commands

All 35, from the Command Palette under AddisDB:

Connections

Command What it does
Refresh Connections Re-read the workspace's connections
Use This Connection Point the editor at this connection
Set Active Connection Pick the active connection
Test Connection Verify it connects
Open Terminal for Connection… Terminal UI, already connected
List Databases on Server Every database on the server
Create Database… Create a new database
Manage Local Database Servers… Start/stop local servers

Running SQL

Command What it does
New Query Open a scratch .sql file
Run Query Run the statement under the cursor (⌘↩ / Ctrl+Enter)
Run Selected SQL Run just the selection
Run SQL File on Connection… Run the file against a chosen connection
Run Script (All Result Sets) Multi-statement script, every result set

Before you run it

Command What it does
What Does This Statement Do? Classify it — kind, read-only or not, warnings
Estimate Query Cost Cost estimate without executing
Count Rows for This Query How many rows it would return
Test SQL in Sandbox Run it against a throwaway copy instead

AI

Command What it does
Generate SQL with AI Prompt → SQL (⌘I / Ctrl+I)
Set AI API Key Store your key; AI is off until you do
AI Privacy Policy for This Database… See and narrow what a model can see

Schema and data

Command What it does
Edit View / Function Definition… Edit a view or function in place
Drop Table… Impact summary, dependents, optional backup
Generate Mock Data Realistic rows for a table
Clear Mock Data Remove generated rows via the ledger

Migrations

Command What it does
ORM: Migration Status What's applied, what's pending
ORM: Review Pending Migrations Each one's SQL plus the safety verdict
ORM: Generate Migration… Generate a new migration

Operations

Command What it does
Show Live Metrics Live server metrics
Show Live Activity (Cancel / Kill) Running queries; cancel or kill them
Show Database Users Who has access
Remove a Database User… Drop a user

Compliance

Command What it does
Scan for Regulated Data (PII/PHI/PCI) Classify columns, with confidence scores

CLI

Command What it does
Update CLI Update the managed addisdb binary
Repair / Reinstall CLI Reinstall it from scratch
Open Website addisdb.com

Keyboard shortcuts

Both apply in .sql files when the editor has focus.

Shortcut Command
⌘↩ / Ctrl+Enter Run Query
⌘I / Ctrl+I Generate SQL with AI

Settings

Setting Default Description
addisdb.cliPath (auto) Path to the addisdb binary. Empty auto-detects the managed install in ~/.addisdb/bin, then addisdb on your PATH.
addisdb.maxRows 1000 Maximum rows fetched per query.
addisdb.autoProvision true Download and install the CLI on startup if missing, and add it to your terminal PATH.
addisdb.mockData.trackInDatabase true Record generated mock-data batches in an in-database ledger (_addisdb_mock_ledger) so they can be listed and cleared later, including from another machine. When off, only the immediate "Remove these rows" action is offered.
addisdb.updateManifestUrl (official) Advanced: override the CLI release manifest URL, for self-hosting or testing.

AddisDB is a product of Cenova Spark LLC. Use is governed by the EULA; see LICENSE.txt.

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
  • Your Privacy Choices
  • Consumer Health Privacy
© 2026 Microsoft