An honest, dated breakdown so you can pick the right tool.
| Feature | TuskDB | the VS Code PostgreSQL extension |
|---|---|---|
| Native macOS app | Yes, Swift & AppKit | No, an extension inside VS Code |
| Platforms | macOS 15+ | Anywhere VS Code runs |
| Database engines | PostgreSQL only | PostgreSQL only |
| Price | $39 one-time | Free |
| Devices per license | 5 Macs, one person | Unlimited (free) |
| Updates | All updates, no renewal | Free |
| Free trial | 14 days, every feature | Free (no purchase) |
| pgvector UI | Yes, find-similar + recall lab | None found (July 2026) |
| Visual EXPLAIN | Yes, computed verdict | Yes, four-view plan visualizer |
| ERD diagram | Yes, incremental canvas | Yes, schema visualizer |
| SSH tunnel | Yes, strict host-key | Not documented (July 2026) |
Verified 23 July 2026 against the Microsoft docs and the ms-ossdata.vscode-pgsql marketplace listing. Microsoft ships often, so check their changelog before relying on any gap below.
Two honest differences. First, pgvector: the extension has no vector UI, so find-similar, the recall lab, and index health are things you would still hand-write SQL for. Second, it is an editor panel, not an app. Database work in TuskDB gets its own window, its own keyboard map, and dense native grids, instead of competing with your code for the same pane. TuskDB also carries SSH tunnelling with strict host-key checking, previewed PK-keyed row edits, and CSV/JSON/SQL export that streams and cancels cleanly.
It is free, from Microsoft, actively maintained, and it works on Windows and Linux as well as macOS. It also runs in Cursor. Its plan visualizer offers four views with per-node cost, rows, timing, and buffers; it has an interactive schema visualizer; and its dashboard surfaces active sessions, blocking trees, and lock activity, so the EXPLAIN, ERD, and activity features are genuinely covered rather than absent. If you already spend your day in VS Code and do not use pgvector, it may be all you need, and there is no honest argument that $39 buys you those three features, because it does not.
Every feature, no account. See if native, Postgres-only, and pgvector-first fits how you work.