Visual filter & sort
Point-and-click filters and drag-to-reorder multi-column sorting. Save any combination as a named view.
v3.5
Smart features — schema diff & migrations, an editable ERD, JSONB & charts
Postgres for humans.
Postgres for humans.
SQL for the curious.
Filter, sort, edit, insert, import and export your tables by point-and-click — while pglens shows you the exact SQL behind every move. Free, open source, and nothing ever leaves your machine.
Star on GitHubEverything you'd open a SQL client for — without typing a query
See it whole — then shape it
Understand any database in seconds
An auto-drawn map of every table and how they connect — hover to spotlight relationships, filter live, and export the whole picture. New in 3.5: the map is editable — drag a handle to add a foreign key, add or retype columns, and pglens turns your edits into reviewable DDL behind the Run button.
Take the tour
A full database client.
A full database client.
None of the SQL.
01
Filter and sort like a spreadsheet
Build a filter from a column, an operator, and a value — no WHERE clause to
remember. Stack multiple sorts by dragging them into the order you want, then save the whole
setup as a named view you can jump back to anytime.
- Type-aware operators: ranges,
LIKE,IN,IS NULL, JSON & array - Drag-to-reorder multi-column sort
- Saved views, shareable by link
| id | customer | status | total |
|---|---|---|---|
| 4821 | Mara Lindqvist | $412.00 | |
| 4820 | Dev Patel | $129.50 | |
| 4817 | Sofia Reyes | $340.00 | |
| 4815 | Jonah Brandt | $118.75 | |
| 4811 | Amara Okafor | $267.20 | |
| 4808 | Felix Tan | $155.00 |
02
Edit and add data, right in the grid
Double-click a cell to change it. pglens picks the right control for the column — a toggle for booleans, a date picker, a validated JSON editor — so you can't write a bad value. Adding a row is a guided form built from your schema.
- Type-aware editing with instant rollback on error
- Insert form that knows your defaults and required fields
- Edits run as safe, pinned-to-primary-key updates
| id | customer | status | total |
|---|---|---|---|
| 4821 | Mara Lindqvist | $412.00 | |
| 4820 | Dev Patel | $129.50 | |
| 4815 | Jonah Brandt |
|
$118.75 |
| 4811 | Amara Okafor | $267.20 | |
| 4808 | Felix Tan | $155.00 |
UPDATE 1 row · 8 ms
03
Follow the data across tables
Click any foreign-key value to slide in the row it points to — no join to write, no second tab to open. Keep clicking to walk the relationships, with breadcrumbs back, or jump to every row that references it.
- One-click foreign-key navigation with breadcrumbs
- "Show all rows where this = that" in one tap
- Edit the referenced row without leaving the panel
04
Prefer SQL? Flip any tab to Advanced Mode
Every no-code action has a Show SQL button, so you see exactly what runs. Want to take over? Flip any tab to Advanced Mode — a schema-aware Monaco editor seeded with the query no-code was about to run, with autocomplete, transactions, timing, and a per-connection history and saved-query library.
- Schema autocomplete,
:nameparams & format-on-save - Transaction mode with
BEGIN/COMMIT/ROLLBACK - Multi-statement results, CSV/JSON export &
EXPLAIN ANALYZEtiming - Per-connection query history & saved queries with folders + tags
SELECT c.name, sum(o.total) AS revenueFROM orders oJOIN customers c ON c.id = o.customer_idGROUP BY c.nameORDER BY revenue DESC LIMIT 5;namerevenue
Mara Lindqvist$4,812.00
Sofia Reyes$3,977.50
Dev Patel$3,214.20
5 rows · 14 ms · auto-commit
05
See what your database is doing, live
The new Operations panel is the tab you open when things are busy or slow:
a live dashboard of active sessions, locks and blocking chains, replication lag and table
sizes — plus a slow-query view over pg_stat_statements, ranked by total or mean
time with an estimated p95. All read-only introspection of Postgres's own stat views.
- Active sessions, blocking chains & connection usage at a glance
- Slowest statements by total / mean time, with estimated p95
- Cancel or terminate a runaway backend in one click
| query | calls | mean | p95 |
|---|---|---|---|
| SELECT … FROM orders JOIN … | 48.2k | 284ms | 512 ms |
| UPDATE sessions SET … | 120k | 62ms | 110 ms |
| SELECT * FROM users WHERE … | 301k | 9ms | 21 ms |
| INSERT INTO events … | 88k | 4ms | 12 ms |
06
Find the slow node — and the missing index
Hand any query to the EXPLAIN visualizer: a tree of the plan with each node's
cost, planned vs. actual rows, and real timing, heat-mapped so the expensive node lights up
instead of its parents. The index assistant reads your catalogs for unused
and duplicate indexes — every DROP INDEX shown for you to review, never run for
you.
- Exclusive (self) time per node, with planner misestimate flagged
- Heatmap by time, rows or cost — plain-English node tooltips
- Unused, duplicate & seq-scan-heavy tables, read-only
Seq Scan orders
Hash Join on customer_id
Index Scan customers_pkey
orders_email_idx
0 scans · 142 MB
DROP INDEX orders_email_idx;
07
Find any table instantly
Hit Cmd/Ctrl+K to open Spotlight and fuzzy-jump to any table or view in any
schema — no scrolling through endless sidebars.
- Fuzzy match across tables & views
- Keyboard-first, opens in a new tab
- Works across every schema, not just
public
◎
ord
⌘K
orderspublic · table
order_itemspublic · table
orders_archivearchive · table
↑↓ navigate · ↵ open in new tab
08
All your databases, one home
Every server lives in a single clean grid. Connections come back when you reopen pglens, and your passwords stay in your OS keychain — never in a config file, never in the cloud.
- Connect by parameters or a connection URL
- SSL modes: Disable, Require, Prefer, Verify CA / Full
- Secrets in your OS keychain, bound to
127.0.0.1
db.internal:5432/main
SSL verify-full
127.0.0.1:5432/dev
connecting…connected
db.staging:5432/main
SSL require
+ New connection
09
Diff two databases, get the migration
Point pglens at two connections — staging and prod, say — and it introspects both and writes the forward and backward migration between them. Statements are ordered for dependency safety and destructive ones flagged red. Like every generator in pglens, nothing runs: the migration opens in the editor, behind the Run button.
- Forward (source→target) and backward migration, both ways
- Dependency-ordered DDL with destructive changes flagged
- Reviewable SQL opened against the connection it targets
+
ALTER TABLE orders ADD COLUMN coupon_id bigint;
+
CREATE INDEX orders_status_idx ON orders (status);
+
ALTER TABLE orders ADD CONSTRAINT fk_coupon FOREIGN KEY …;
!
ALTER TABLE customers DROP COLUMN legacy_ref;
opens in the editor — behind the Run
button
10
Chart any result, no re-query
Flip any query result or table page to the Chart view and pglens plots the rows already in hand. It reads your columns — numeric, temporal, categorical — to suggest a line, bar or scatter with sensible axes, and you can override the type or either axis.
- Line, bar & scatter from the rows you already loaded
- Auto-suggested chart type and axes from column types
- No second round trip — it plots what's on screen
LineBarScatter
from current rows
Why pglens
Built to do the work,
Built to do the work,
not just show the rows.
Other clients make you write SQL for everything, or lock the good parts behind a subscription. pglens does it all by point-and-click — for free.
Inline editing & insert
Edit cells with type-aware controls and add rows through a schema-built form. No bad values, no UPDATE statements.
Follow foreign keys
Click a value to open the row it references, walk relationships with breadcrumbs, and edit along the way.
One-click aggregations
Count, sum, avg, min, max, stddev and distinct counts pinned to the bottom of every grid — computed on the active filter.
Export & import
Export tables to CSV, JSON or SQL respecting your filter and columns. Import CSV with column mapping and conflict handling.
Schema visualizer
Auto-laid-out relationship graph with hover spotlight, live filtering, minimap, and export to SVG or Mermaid ER.
Show SQL & Advanced Mode
Every no-code action reveals the exact query it runs — and any tab flips to a schema-aware Monaco editor with transactions, timing, history & saved queries.
Operations dashboard
Live activity over Postgres's stat views — sessions, locks, replication, sizes — plus a
slow-query view on pg_stat_statements. Cancel a backend in a click.
EXPLAIN & index advisor
A heat-mapped EXPLAIN plan tree that lights up the costly node, and read-only advice on
unused, duplicate & seq-scan-heavy indexes — every DROP shown to review.
Schema diff & migrations
Diff two connected databases and generate a forward and backward migration — dependency-ordered, destructive changes flagged, opened in the editor to review.
Visual ERD editor
Drag a handle to add a foreign key, add or retype columns, drop them. Edits become reviewable DDL built from structured ops — never SQL fragments from the browser.
JSONB explorer
Sample a jsonb column to infer its key set and types, then click a key to add a
native ->>, @> or ? predicate to the filter
bar.
Extensions panel
Browse available extensions with installed vs. default version, then install or drop with
one click — RESTRICT, never CASCADE, DDL shown for review.
Charts
Plot any result or table page — line, bar or scatter — from the rows already loaded, with chart type and axes auto-suggested from your columns. No re-query.
Private by default
Runs on your machine, bound to 127.0.0.1, secrets in the OS keychain. No
account, no telemetry, nothing in the cloud.
FAQ
Questions, answered.
Do I need to know SQL?
No. Filtering, sorting, editing, inserting, aggregating, exporting and importing are all point-and-click. SQL is optional — and when you're curious, "Show SQL" reveals exactly what each action runs.
Can I write my own SQL when I want to?
Yes — flip any tab to Advanced Mode for a schema-aware Monaco editor with
autocomplete, :name parameters, format-on-save, transaction mode
(BEGIN/COMMIT/ROLLBACK), multi-statement results
with EXPLAIN ANALYZE timing, and a per-connection query history plus a
saved-query library with folders and tags. It opens seeded with the query no-code was
about to run, so you start from a working query, not a blank page.
How is pglens different from pgAdmin, DBeaver or TablePlus?
Most clients are SQL-first — you write queries to do anything real, or pay for the no-code parts. pglens flips that: filter, edit, insert, follow foreign keys, aggregate and import/export without typing SQL, with the generated query always visible. Free, open source, and runs locally.
Is editing my data safe?
Yes. No-code actions never send raw SQL from the browser — the server turns your choices into parameterized queries. Edits are pinned to the primary key, type-validated, applied optimistically, and rolled back if the database rejects them.
Can pglens help when the database is slow?
Yes. The Operations panel shows live activity (active sessions, locks and
blocking chains, replication, sizes), a slow-query view over
pg_stat_statements, an EXPLAIN plan visualizer that heat-maps
the costly node, and an index assistant for unused and duplicate indexes. It's all
read-only introspection of Postgres's own catalogs — the only SQL it generates is a
DROP INDEX shown for you to review, never run for you.
Can pglens generate migrations or edit my schema?
Yes. Diff & migrate compares two connected databases and produces a forward and backward migration; the visual ERD editor lets you add foreign keys, add or retype columns and drop them. Both turn your changes into dependency-ordered DDL with destructive statements flagged — and, like the index assistant, pglens never runs it. The SQL opens in the editor, behind the Run button, built server-side from structured operations rather than SQL typed in the browser.
Is pglens free?
Yes. pglens is open source and MIT licensed. No account, no telemetry, no paywall.
Is my data sent anywhere?
No. pglens is local-first — it binds to 127.0.0.1 only, routes are
token-gated, and passwords are stored in your OS keychain. Nothing leaves your
machine.
Which platforms and Postgres versions are supported?
macOS, Linux and Windows, with any modern PostgreSQL server. One install command sets up Node.js and every dependency. Connect by parameters or URL, pick any schema, choose your SSL mode.
Get started
Stop writing SELECT *